Priority color codes: 1 2 3 4 5
# Type Status Created By Changed Assigned Svr Pri Title Opis  
90 new active 2002 Aug anonymous 2002 Aug drh 3 2 Handle different projects in one CVSROOT When more than one project is located under one CVSROOT it is difficult (or impossible) to create an Milestone for an special project. Remark: Yes, I know how to it from the command line. edit
99 new active 2002 Aug anonymous 2002 Aug   4 4 Cvstrac home or index page not found when project is omitted Cvstrac displays this error page when no project is included in the request URL: ----- *Not Found* Page not found: / ----- We installed cvstrac as a CGI script in the admin directory of our web server. Cvstrac responds correctly to requests like /admin/cvstrac/project1 and /admin/cvstrac/project2. When passed the request /admin/cvstrac it displays the error page. I recommend that cvstrac respond with either an index of available projects or a static HTML page. edit
158 new active 2002 Dec anonymous 2002 Dec drh 4 4 Make ticket page a Wiki page with generalised meta-formatting If the Wiki supported keywords like {SQL: command }, then it would be possible to create Wiki pages that interface with the database. In this scheme, a ticket page might be written in the Wiki like (I don't know any SQL yet): Ticket {SQL: ITEM.number }: ticket dependencies and estimated duration {SQL: ITEM.description } Remarks: {SQL: ITEM.remarks } And editing the page will bring up edit blocks which feed the data back into the database, like the Create New Ticket page. The same idea could apply to reports, so the generated report page could also have an "edit" tag, which brings up the same information in edit blocks for editing the database contents. This would let me perform tasks like create a report for all entries older than 4 weeks along with their priority levels, click [Edit] and directly modify all their priority levels. This would mean that Ticket pages are just Wiki pages with meta structural information, and all Wiki pages can extract and modify SQL database information. This scheme also handles the scenario in the TODO item: * Text above ==== may only be editted by the document owner edit
178 event active 2003 Feb anonymous 2003 Feb drh 3 4 Old or legacy rcs (not sure) is unable to handle 'date <= date' Import of CVS history generates loads of errors: No revision information found in rlog output: rlog '-d1999-06-01 13:48:14<=2003-02-12 09:21:14' '/daily/CVSROOT/CVSROOT/loginfo,v' 2>/dev/null; Runnin rlog manually, I got: rlog error: can't parse date/time: =2003-02-12 09:21:14 rlog aborted Modifying '<=' to '<' at the appropriate line in history.c solves the problem for me. Not sure if it's suitable as general solution. edit
175 new new 2003 Feb anonymous 2003 Feb drh 3 4 A second CVSTrac logs out the first one. I use CVSTrac for two projects. I usually have two browser tabs open, each showing the CVSTrac timeline for one of the projects. After I've logged in for the second CVSTrac, the a refresh on the first one goes to the login page. Logging in again for the first one causes the second one to be logged out. It would be nice if I could be logged in on both projects simultaneously (some cookie manipulation required?). edit
180 new active 2003 Feb anonymous 2003 Feb drh 3 3 Larger text fields in ticket editor I'd like to be able to set, perhaps in the setup page, the number of lines for the description and remarks fields. edit
181 event active 2003 Feb anonymous 2003 Feb drh 3 3 No <nobr> in ticket editor In tktedit, there are <nobr> tags between all the popup menus and fields (status, type,...,contact). I'd like these to be removed so that I don't have to make the browser window really wide to see them all. ---- I'm having the most trouble with Safari (Apple's new KHTML browser). Even though the Associated Check-ins text field is 70 wide (why?), it never wants to put a break before it. The W3C Validator said that <nobr> was invalid. When I remove those tags, it renders fine in Safari. I can attach some screenshots if you want, but for some reason it's not letting me (even thought it said I would be given the opportunity to add one when I created the ticket). edit
191 event active 2003 Mar anonymous 2003 Mar drh 3 1 User keeps getting logged out One of my developers keeps getting logged out. He says a typical session goes like this: Yip. Cookies enabled. I just tried it under netscape as well. 1) Login 2) Wiki 3) edit (or was it note, edit?) 4) Now logged in as anon. not Andre Same thing happens with making or editing tickets. edit
210 new active 2003 Apr anonymous 2003 Apr   3 4 what about numbers in CamelCase words ? I'm starting to use Wiki on a project and I need to create pages with numbers in name (i.e. SubProj101, SubProj102 etc.).{linebreak} The name is not converted to hyperlink, it appears as the presence of digits violates the coding, but unfortunately a lot of names (especially technical-related) use digits. it could be enough to make a little change in format.c: = Actual = int is_wiki_name(const char *z){ int i; int nCap = 0; if( !isupper(z[0]) ) return 0; for(i=0; z[i]; i++){ if( isupper(z[i]) ){ if( !islower(z[i+1]) ) return 0; nCap++; }else if( !islower(z[i]) ){ break; } } return nCap>=2 && is_eow(&z[i],0) ? i : 0; } = Changed = int is_wiki_name(const char *z){ int i; int nCap = 0; if( !isupper(z[0]) ) return 0; for(i=0; z[i]; i++){ if( isupper(z[i]) ){ if( !(islower(z[i+1]) || isdigit(z[i+1]) ) ) return 0; // <<< nCap++; }else if( !islower(z[i]) ){ break; } } return nCap>=2 && is_eow(&z[i],0) ? i : 0; } Without changes in wiki_expand_name (wiki.c) the expansion will be: SubProj101 -> Sub Proj101, It should be acceptable.{linebreak} BTW: I included code fragments in <pre> and </pre> but it didn't worked. _Indent by two or more spaces to get verbatim text_ edit
213 new active 2003 May anonymous 2003 May drh 5 3 Readying cvstrac for non-english alphabets The attached diff to cgi.c will allow use of danish and swedish (and probably all other) under netscape. edit
164 doc active 2003 Jan anonymous 2003 Jun drh 5 3 Yet another multiple project per repository hack I've locally tweaked the 1.1.2 source code so that once read into a temporary 'fc' table, entries get deleted through an SQL query like 'DELETE FROM fc WHERE <config::module_delete_where>'. I've stored an SQL expression in the config table under the 'module_delete_where' identifier. Admitted, this is a quick and dirty hack but I definitely needed this feature as I have multiple projects sharing common modules in my unique repository. FYI, the hack is about 'setup.c' and 'history.c' files. The '.0' files are those of v1.1.2 release. Hope this helps. > diff -u setup.c.0 setup.c --- setup.c.0 2003-01-23 18:20:56.504453000 +0100 +++ setup.c 2003-01-21 17:32:42.468555000 +0100 @@ -117,6 +117,7 @@ }; const char *zRoot, *zOldRoot; const char *zModule, *zOldModule; + const char *zModuleDeleteWhere, *zOldModuleDeleteWhere; /* The user must be the setup user in order to see ** this screen. @@ -155,6 +156,21 @@ } /* + ** The "x" query parameter is the expression matching the name - + ** actually what shall not be deleted - of the module within the + ** CVS repository that this CVSTrac instance is suppose to track. + ** Change it if it has changed. + */ + zOldModuleDeleteWhere = db_config("module_delete_where",""); + zModuleDeleteWhere = P("x"); + if( zModuleDeleteWhere && strcmp(zOldModuleDeleteWhere,zModuleDeleteWhere)!=0 ){ + db_execute("REPLACE INTO config(name,value) VALUES('module_delete_where','%q');", + zModuleDeleteWhere); + zOldModuleDeleteWhere = zModuleDeleteWhere; + db_config(0,0); + } + + /* ** The "rrh" query parameter is present if the user presses the ** "Reread Revision History" button. This causes the CVSROOT/history ** file to be reread. Do this with caution as it erases any edits @@ -219,15 +235,23 @@ @ (for example, if you want CVSTrac to see only one module in a @ repository that contains many unrelated modules) then @ enter a pathname prefix for the files you want to see in the - @ second entry box.</p> + @ second entry box. If you want to ignore a subset of the files + @ contained in the CVS repository (for example, if you want CVSTrac to see + @ only some modules in a repository that contains many unrelated modules) + @ then enter an SQL expression of the modules to be filtered out while + @ parsing the history file.</p> @ <p><table> @ <tr> @ <td align="right">CVS repository:</td> - @ <td><input type="text" name="r" size="40" value="%h(zOldRoot)"></td> + @ <td><input type="text" name="r" size="80" value="%h(zOldRoot)"></td> @ </tr> @ <tr> @ <td align="right">Module prefix:</td> - @ <td><input type="text" name="m" size="40" value="%h(zOldModule)"></td> + @ <td><input type="text" name="m" size="80" value="%h(zOldModule)"></td> + @ </tr> + @ <tr> + @ <td align="right">Module 'delete where' expression:</td> + @ <td><input type="text" name="x" size="80" value="%h(zOldModuleDeleteWhere)"></td> @ </tr> @ </table><br> @ <input type="submit" value="Submit"> > diff -u history.c.0 history.c --- history.c.0 2003-01-23 18:21:06.814517000 +0100 +++ history.c 2003-01-21 14:02:56.432465000 +0100 @@ -426,7 +426,20 @@ zModule ); } - + + /* + ** If the "module_delete_where" configuration parameter exists and is not + ** an empty string, then delete from the FC table all records dealing with + ** files that are not matching the fetched expression. + */ + zModule = db_config("module_delete_where", 0); + if( zModule && zModule[0] ){ + db_execute( + "DELETE FROM fc WHERE %s", + zModule + ); + } + /* ** Extract delta comments from all files that have changed. */ edit
254 new active 2003 Oct anonymous 2003 Oct   5 4 Class & file names that look like CamelCase In our java development project, we often have docs and commit logs that refer to classes & source files that look like CamelCase -- e.g. ChangeLog, HashCache, MailTransferAgent. I've found myself clicking on these links in the vain hope they would take me to the source file or class in question. I'm wondering if you could have the CamelCase link directly to the source file; right now I usually click the link, and then edit the page to have a link to the relevant source file, but I like to use the "Contents" page to surf around, and I dont like having clutter in there. edit
120 new active 2002 Sep anonymous 2003 Oct drh 4 3 make -j fails make -j fails with something _remotely_ similar to "I am too dumb to make page_index.h" To fix, apply this patch (uudecode and gunzip it) or change the "headers:" target in main.mk to "page_index.h headers:" (works with GNU make at least, not sure about BSD make though) begin 644 main.mk.diff.gz M'XL("'J>>CT"`VUA:6XN;6LN9&EF9@"-4<%2@S`4/).O>`</.A@$I5HY.#C> M/;3>.RF$DH%`)PE6_]X\H"4P/<@A[^W.9H?LYJ(H@'9`511$$4@FFD!6A%)Z MWKT5;/D1'L/P$:)5$K\F3\]`0_MY>(7XOG^1KB=I&";Q.HGB04K2%&@4QO<O MX`\C30EX2@(MP"C6Z)H9;GTJ7G*6<Z5!5J+)^0\<V8'O^C4H89Q&UC#J+BY[ MU9XTMYKL(/!LI6P;N^1[>Q2MDLS8I13:M.H7K:0UGBRA;@\"]?U;[,A7]M2< MJ:SL%],=[31<HXT16<6'1?):-&C4::[L^!;\9,=)5&(<HA%62H#0\9\3[]I# M;VZ_-N^?V]UV\W%'_-FS_W\/O.#!%8VQ[((L<0-"[.;4XRDQA$YN")<)(K<( M$ZE9JCWAQHN$FS/B*>\>.;$CGA<P,%,3B&>-],2RFX&<]824TQ="IS>$B_[. ;U+E+C/E*_O#F]F9EINVR2W_D#V*6"K)N`P`` ` end edit
255 new active 2003 Oct anonymous 2003 Oct   5 4 Dumping javadocs into Wiki It would be nice to have a tool to dump javadocs (and I know Python has a similar system which would also benefit) into so your commit comments and wiki pages can easily link to that documentation. I'm not sure whether it would be better to change the parser to treat java.lang.String (and perl::CPAN, and the python version, etc.) as equivalent to JavaLangString, or to just add a suffix to the class names, like StringDoc; obviously the first would allow the javadocs themselves to be parsable as a wiki. The first is unappealing because of its language-specificness and so on, but of course its more "handy" for the users. The second could be done in anyway -- all you really need is a tool that allows you to add an arbitrary file as an arbitrary wiki page, and let the user write his own script to put which file where. Test text... edit
263 event active 2003 Nov anonymous 2003 Nov drh 4 4 CGI problem with Opera browser I'm on a Linux system and use the latest Opera browser (version 7.21). I was trying to edit a page on the SQLite Wiki. When I clicked on "Preview Your Changes" I got the editing page back, not the preview. I tried the same thing on Mozilla. The preview worked fine. I submitted the edit and that worked fine too. Then, back in Opera, I went to the cvstrac wiki to create this ticket and had the same problem. Clicking on "Preview Your Changes" just brought the edit page back. Finally, I went back to Mozilla where I've been able to successfully create this ticket. edit
179 new active 2003 Feb anonymous 2003 Nov drh 3 4 Making log message changes in the underlying ,v file Right now, editing the log message for a checkin from within CVSTrac does not modify the underlying CVS ,v files. To reflect any such changes in the underlying files manually is painful, since there may be multiple files in the checkin, and each could have a different revision number. It would be useful if CVSTrac could (as an option) modify the log message in the underlying ,v files (using the `cvs admin' command, or in some other way). edit
294 doc active 2004 Mar anonymous 2004 Mar anonymous 1 1 Documentation on the CVS Repository CVSTRac doesnt recognize the CVS Repository even if you do a commit after an import. You must actually manually add each file like this cvs -d $CVSROOT add [file] then CVSTrac will display the CVS directory that this file resides in. edit
299 new active 2004 Mar anonymous 2004 Mar drh 3 2 Add enums for severity and priority The numeric constants (1-5) for severity and priority are a source of confusions for our users - other systems that we have are using different numbering systems. It will help if the "enum" table will be extended to allow the assignment of labels to those levels. There is no need to change the actual coding - the 1-5 level are more than enough. The enums should be edited by the 'setup' user, and should be used for ticket entry and editing. edit
302 new active 2004 Mar anonymous 2004 Mar drh 5 5 Allow queries to be embedded in other pages. It would be very useful to be able to add a tag to reports to include them in other pages. A combo box on the reports page could list destinations where the report could be embedded (ie: ticket page, milestone page, ticket edit page, home page, etc). By enabling select entries the user could embed the report in different pages. This could be used to provide a very simple mechanism to track dependancies. I have a report that lists all tickets that depend on a specified ticket (see attachment). By embedding this query in the ticket overview page I could follow a link to any dependent tickets. A similar report on the milestones page would allow all tickets that depend on a milestone to be listed. It might be useful to embed the "Ticket Counts" report on the main page. etc. One issue I don't know how to resolve would be passing the page identifier to the query. Each embedding location could define a key (or keys) it would make available to the query, but I don't know how it would be incorporated into the query. I assume something similar to the aux() mechanism that is currently used would work (I'm not much of an SQL guy). So a report embedded in the ticket page would have the ticket id supplied, and a report embedded in the timeline page would have the commit id supplied. edit
303 new active 2004 Mar anonymous 2004 Mar drh 4 5 Milestone and ticket hotlists. Referencing tickets by id can be cumbersome when composing a new ticket. Typically it requires a separate browser tab to search the database. Since some tickets and milestones are referenced more than others it would make sense to make their ids and summaries available when composing a new ticket. If a new column were added to the milestones and tickets tables called hot, by checkmarking the hot column on the edit page for the ticket they could be added to a 'hotlist'. Whenever a field was presented that accepted a list of dependancies (such as the depends on field and the associated check-ins field on the edit ticket page a combo box next to the field could list all 'hotlist' ticket numbers and check-ins. When the page was submitted the entries in the combo box would be merged with the input in the text field. I think this would be particularly useful when designing with milestones, since all milestone check-ins could be marked "hot" and then when reviewing tickets I could target them for a particular release by selecting the version number from the list. Combined with suggestion #302 this would make using the timeline and milestones to plan the future direction of the project much more effective. edit
304 new active 2004 Mar anonymous 2004 Mar drh 5 5 Option to pipe reports through an external filter. Currently reports can be displayed formatted into a table or as raw data. A third option to provide an external filter script (much like the external diff programs currently supported) to filter the report would offer several new configuration possibilities to the user. The raw data would be piped through the filter and the result returned. Two columns added to the reports table would control this. The first would contain the script command to run. The second would contain the mime type of the returned data. When embedding the result in a page the tag would be based on the mime type of the result. If it were of type text/html (or blank) it would be embedded directly in the page. If it were of type image/* it would be embeded by an image link, if it were type text/plain or ?? it would be referenced by an a anchor link. This would allow a lot of additional functionality. For instance a report that extracted ticket dependancies could be piped into a perl script that generated a dot file which in turn would be piped into dot to generate a dependancy graph. I think it would also have been possible to implement the rss feed this way. Other feeds for other systems could be generated in a similar way. Even the current report tables could be implemented as a perl script (this wouldn't make much sense however since it would then be impossible to install cvstrac as a single file). One security issue is that allowing users to configure an arbitrary filter to run on the server is not very safe :-). Such functionality would have to be limited to the server admin. edit
305 new active 2004 Apr anonymous 2004 Apr lansil 1 1 css customization Is there scope for the site to be customized using css ? edit
308 new active 2004 Apr anonymous 2004 Apr drh 5 3 Shaded "Initial Checkins" on Branches in the Timeline Thanks 4 that really nice tool... But it could even be better, when initial checkins on branches would be shaded like checkins on existing files in branches... That would be very helpfull... Thx in advance....John edit
317 new active 2004 May anonymous 2004 May drh 4 3 provide simple test mechanism for email notification I have throughly tested a CVSTrac mail notification script I downloaded and customized (going so far as to run it with fake input as the webserver user. It seems to work fine, except I am receiving no CVSTrac mails, and logging does not appear to work (next bug). It may be relevant to consider actually integrating email notification into the next release, or at least packaging scripts that do mailing which are known to work under a variety of circumstances. edit
324 new active 2004 May anonymous 2004 May anonymous 3 4 viewcvs/cvsweb compatibility of urls Cvstrac provides a browse capability which is (except for tickets) a strict subset of that provided by viewcvs (most common) and cvsweb (older) products. Many cvs clients assume that viewcvs or cvsweb is available and send GET-style urls to the appropriate web browser. To lessen the need to install viewcvs/cvsweb, an even larger subset of their REST architecture could be implemented. Cvstrac already does much of what cvsweb_etc does: *: directory browsing - cvstrac uses /dir?d=/path . To emulate cvsweb_etc we could use /viewcvs/path *: file download- cvstrac uses /rlog?f=/path.... cvsweb_etc emulation could be /viewcvs/path also. *: file download and versions - cvstrac uses /getfile... cvsweb_etc have either "checkout"/path?rev=.. or ~checkout~/path?rev=.. *: diff-cvstrac uses /filediff?.. cvsweb does /path.diff?r1=..., viewcvs does /path?r1=... Implementation could be simple redirects from a new function hanging off /viewcvs. /path maps to g.zExtra (as in getfile); deciding which to redirect would be based on an examination of /path and a few looks at parameters P(r1), etc. I will look into this. --Derek edit
325 new active 2004 May anonymous 2004 May drh 3 2 Make wiki diff configurable The diff output for wiki entries should be configurable - both in terms of diff format (unified format preferred), and the ability to pipe to an external process so pretty output can be displayed. The stuff you already have implemented for the cvs external diff processes would work great for the wiki diff processes. (Note that the external diff config doesn't make clear that it only applies to the cvs part and not the wiki part.) edit
329 new active 2004 May anonymous 2004 May drh 5 4 Request wiki calendar It would be really great if the wiki had a calendar. I am thinking of something where developers could show what meetings they are going to, when they will be out of the office, that sort of thing. Even nicer would be some sort of aggregation to show group calendars. This sort of thing is actually quite hard with wikis since you either end up with lots of pages (eg one per day) or some sort of horrible formatting nightmare. I love the Timeline in CVS trac so hopefully it could be part of that as well. edit
332 new active 2004 Jun anonymous 2004 Jun drh 4 4 Shortcuts to favourite reports in the cvstrac page header First: what a great thing cvstrac is. Just right for what I need. Nice to look at as well. Thanks a lot for it! Here's my (minor) quibble: I have a few report types I often use and I find myself having to go to the reports page and then click the report I like there. This always feels like a little detour... ;-) If each report had a check-box 'show in favourites', those reports (with the box checked) could be displayed in an area just below the blue page header box (or in an extra line inside the box). That would allow quick access to these reports. Saves one page-load every time (the load of the 'reports' page). edit
336 new active 2004 Aug anonymous 2004 Aug drh 1 1 Missing Fields in CVS Trac PROBLEM Not enough multi line fields SYMPTOM There is only 1 multi line field in the screen 'Create a new Ticket' SOLUTION In order to keep descriptions simple I'd like to have a set of different multi line fields: - Symptom - (Requested)Solution - Steps to Reproduce COMMENT Many people who report an issue make it 1 long story. Using separate fields for each part will force them to clearly state their point. The single line possibilities in 'User-Defined Fields' are not an option. edit
313 doc active 2004 Apr anonymous 2004 Oct   2 2 Date shown in timeline is wrong When looking at timeline and chgview, cvstrac is off by a day, here's a sample check-in: Date: 2004-Apr-20 17:40:23 (local) 2004-Apr-20 21:40:23 (UTC) [snip] --- LDAP.pm 2004/03/22 17:02:53 1.45 +++ LDAP.pm 2004/04/21 21:40:23 1.45.2.1 Notice the 1 day difference between the date reported by chngview 2004-Apr-20 (UTC) and diff 2004/04/21, same problem shows in timeline, all dates are off by 1. I am pretty sure this is not a bug but something to do with our configuration, but can't figure out what and where to configure this. Thanks for any insight. edit
369 doc active 2005 Jan anonymous 2005 Jan drh 1 1 Studying the source code Could anybody give me some background on the design philosophy behind cvstrac? I like the software, and I like the idea of writing native CGI applications. So, I started digging in the cvstrac sources to get inspiration. I understand individual lines, but I fail to understand the big picture. What is the guiding principle behind the subdivision in separate files. Usually, you find simple guiding principles, such as: one source file corresponds to one web page (typical in php, asp, asp.net, ...) or corresponds to one abstract data type (such as one table and its related methods, or one memory struct and its related methods, ...), or is otherwise organised around another particular guiding principle, which tremendously facilitates locating particular logic. It turns a source tree into an addressable grid, where you can almost calculate where some logic must be located, and where you can make changes, if needed. What is the guiding principle underlying cvstrac? edit
370 new active 2005 Feb anonymous 2005 Feb drh 1 1 Why not to scan the cvs directory to get the reponsitory? Why not scan the cvs directory to get the reponsitory? I setup the reponsitory of cvstrace, but the history file don't includes all informations of the modules, so I Only can get the Reponsitories which is in history file. edit
373 new active 2005 Feb anonymous 2005 Feb drh 2 2 Add ability to create templates for new tickets It would be helpful if we have the ability to reuse mutliple templates to open new tickets. Expected: 1: Should be able to save a template with default values set to different parameters. 2: Reuse these templates to open different types of tickets edit
376 new active 2005 Mar anonymous 2005 Mar drh 1 1 create a .pid when run in server mode - crontab friendly I am running CVSTrac as a standalone server. It would be nice, when standalone server is launched, if it created a pid file. That way I could easily check from crontab is CVSTrac is still running and restart it if not. edit
386 new active 2005 Mar anonymous 2005 Mar drh 5 4 "Show Milestones" does not show milestones generated by cvs rtag When browsing the source code of a project and clicking "Show Milestones" only milestones that were created with cvstrac show up. Milestones generated by "cvs rtag" do not. Additionally, when viewing the timeline with "Divide timeline by milestones" checked, again, only milestones entered manually into cvstrac show up in the pink boxes. It would be nice if the default behavior when "Divide timeline by milestones" is checked was to list the "cvs rtag" generated milestones without ever having to manually add a "Release" milestone in cvstrac. edit
396 new active 2005 Apr cpb 2005 Apr cpb 5 4 owner should be able to lock tickets Ticket owner/assignee/admin should be able to lock a ticket against changes, similar to how wiki locking works. edit
397 new active 2005 Apr cpb 2005 Apr cpb 5 4 ticket link info should contain status #392 gives title information and #52 indicates at a glance whether or not the ticket is active. However, _why_ it's not active would be extremely valuable... i.e. was it fixed or deferred? Is a ticket new or actively being worked on? If we're looking at the ticket status anyways, we might as well present it to the user... edit
428 event active 2005 Jun anonymous 2005 Jun drh 1 2 cvstrac does not see modules, which are created via pserver I today installed cvstrac running from inetd under FreeBSD 4.11 from the port. I did setup the cvsroot path and left the module field blank, to see all modules. When browsing cvs from cvstrac I see some modules missing. When checking CVSROOT/history I saw that the missing modules were created remotely via pserver access using cervisia frontend. But I can see other remotely created modules. There are 88 modules in our repository. edit
361 new active 2004 Dec anonymous 2005 Aug drh 4 3 localisation <html>Hi,<br/><br/> I am interested in a german localisation of cvstrac. Therefore would do the translationwork!<br/><br/> <b>It would be nice to see plain text sourced out to an file</b></html> edit
436 new active 2005 Aug cpb 2005 Aug cpb 5 4 add filters to browse directory view The ability to filter a directory view in a browse page with "globs" would be real handy when looking for something specific. All you'd need is just a textfield entry at the top of every directory "/dir?" page. edit
448 new active 2005 Aug anonymous 2005 Aug cpb 1 1 import of svnserve users Attached diff enables _svntrac_ to import users from _svnserve_ users file. This is not equivalent to reading of CVSROOT/passwd since some installations of _Subversion_ don't use _svnserve_ and hence don't have relevant users file. This is not the best solution, but ATM I have no plans to add support for importing users from _{quote: mod_dav_svn}_ . _svntrac_ supports only import of users, and unlike _cvstrac_ , it can't update _svnserve_ users file based on the contents of _svntrac_ db. This limitation is imposed by the fact that passwords are stored encrypted in _svntrac_ db, while they are stored as plain text in _svnserve_ users file. Only way, AFAICT, to support it would be to store plain text passwords in db, but I guess this goes against _cvstrac_ design goals?{linebreak} Though it would be nice to have this "user export" feature since AFAIK there is no way to manage _svnserve_ users other then to edit users file directly. edit
450 new active 2005 Aug cpb 2005 Aug cpb 4 4 append remarks could be smarter I'm starting to see where remark handling could be enhanced to have some concept of "replying" to a particular comment. Nothing as sophisticated as message threading, but maybe just filling in with the original remarks text. Something to help out in building context... This would require some way for the ticket formatter to distinguish between individual remark sections... edit
243 new active 2003 Aug anonymous 2005 Aug drh 5 4 Allow more customization for the header and footer I've had a need to customize the header a bit more, keeping a company logo on the top bar, not above it. Attached to this report is a patch that allows slightly more customization. I think it would be valuable for general consumption, as I can imagine this a common need. edit
459 doc active 2005 Aug cpb 2005 Aug cpb 5 3 Fix documentation/UI to make more sense with SVN There's a lot of references around to things like _CVSROOT/history_. edit
476 new active 2005 Sep cpb 2005 Sep cpb 5 4 add support for GIT http://git.or.cz/ There's two main reasons for this. 1: Documentation. SvnTrac a poor guiding example for adding a new SCM because we were making things up as we went along. Documenting _that_ process is just going to introduce a bunch of noise. 2: Proof of concept. GIT is quite far from the CVS model. Getting it working is a good test of the whole ScmTrac concept. It appears that it shouldn't be too difficult to get things working. edit
475 doc active 2005 Sep drh 2005 Sep drh 1 1 Add wysiwyg editing of wiki and tickets. See http://www.wikiwyg.net/ edit
479 new active 2005 Sep anonymous 2005 Sep drh 4 4 Source code files should have remarks/attachments Source code should allow for remarks and attachments to be associated with a particular file. While this sort of thing could be put into the code itself, it generally is not, since it may bother the writer. And even if it were, only programmers would be able to do so. In general, there are 5 types of things in cvstrac: tickets, remarks, attachments, source code files, and wiki pages. Tickets seem to be first-class objects (#123 anywhere refers to a ticket), and Wiki pages seem to be becoming that as well. Remarks and atachments are associated only with a particular ticket or wiki page. But source code seems completely unrelated to these. Does the viewer highlight Ticket references or WikiNames in the code? It probably should not because it may make a mistake. At the same time, associating external documentation with a source file seems valuable. (One could conceive of this becoming a full-featured Literate Programming methodology). I would think that the implementation of this would be to have options on the rlog view such as: the number of attachments, number of remarks, number of tickets against a particular file, a link to the annotate/blame view (Ticket #183), edit
484 new active 2005 Sep anonymous 2005 Sep drh 5 2 Please increase the rows and cols of the textarea in ticket page rows="20" cols="100" is preferred. Or add function let user set them by himself. Because when I write more lines, the small textarea makes me feel as look through a small window to the outsider. thanks. edit
478 new active 2005 Sep cpb 2005 Sep cpb 5 4 should be able to browse tree "snapshots" Browse as currently implemented shows, basically, a union of all sets of files ever found in the repository. It would be nice if there was an option to browse the tree as a snapshot at any particular changeset. That way you don't see the deleted stuff and you don't see stuff added later. edit
488 new active 2005 Sep anonymous 2005 Sep cpb 4 4 Enhancing derivedfrom (and providing backward compatibility) Currently, we used derivedfrom to record external references as to where where a ticket was derived from (e.g. change request - CRxxx etc) as well as references to other tickets (1.1.4 didn't really state otherwise) . This functionality has now changed in 1.2 . Additionally, derivedfrom can currently only reference one other ticket. I am requesting comments on the suitability of the following changes for inclusion in CVStrac (I'll do the work - however, I don't want to have to maintain my own version and will just use an 'extra' field if the following is not acceptable). *: Add additional xref table for derivedfrom tickets references. *: Preserve derivedfrom field as entered in webform for further display. *: Scan derived from field for any number or #number which hasn't been prefixed or sufixed with letters. Store these references in the new xref table. *: Use xref table to look up derived tickets for use in the ticket view edit
399 new active 2005 Apr cpb 2005 Oct chorlya 5 4 add option to limit timeline size by number of lines Alternative to number of days. Also makes projects that periodically stop look more active ;) edit
503 doc active 2005 Nov anonymous 2005 Nov drh 3 3 A Better "FormattingWikiPages" + 2 wiki bugs I have identified two outright bugs in cvstrac's wiki (probably already documented somewhere else; if not, someone on the cvstrac team should probably ticketize them and assign them to the right people): 1: If the entire link: statement will not fit on one line inside the wiki text editing box, the link command is ignored. 2: * followed by _ then some text then _ followed by * (and vice-versa) to make text both bold and italic don't work, despite everything else being nestable (e.g. *'s around a "quote:" block); the more interior of the two markups is simply interpreted as part of the "word" the exterior markup is affecting. A third issue (a misfeature) is that use of html and /html to insert arbitrary HTML code introduces a bogus, worse-than-useless linebreak, which makes the html wikicommand only usuable on large blocks of stuff, instead of inline in the middle of a sentence of whatever. That said, I was in the process of modifying my local copy of FormattingWikiPages to deal with problem no. 1 above, and just decided to overhaul the entire thing, as a way of giving back. There were several details in the editing hints at the bottom of the page when editing any wiki page, that were not in the FormattingWikiPages documents, at least one typo, inconsistent use of bold to highlight wiki markup commands, unclear phrasing in several places, and a number of other issues. So, here is my take on it, use any of it that you like. I did implement big "WARNING:" bits about bugs and other problems, that of course should come out of this version of the wikitext if you want to use it, and the issues get resolved by then. ;-) PS: If posting the wikitext into this ticket mangles it in some way, drop me a line and I'll e-mail you a copy. The preview I just did seems to indicate that cvstrack will render all the wiki commands even inside a ticket, so you can't see the wikisource. I should probably e-mail it. ***************************************************************************************** **Paragraphs** Use *one or more blank lines* to separate paragraphs. You can arbitrarily wrap text within a paragraph (or not bother to wrap it at all), _except_ that {quote:{link:}} commands cannot line wrap; see below for more details. **Indentation of Code** If the first line of a paragraph begins with a *tab* or with *two or more spaces*, then that whole paragraph is shown verbatim (that is, within <pre>...</pre> markup in HTML, including the constant-width font with all spacing and line breaks preserved) The formatting rules below _do not apply_ to verbatim paragraphs (other than {quote:<html>}, as explained next). See "Lists", below, for how to indent without <pre> behavior. **HTML** You can insert arbitrary HTML into a wiki page, by surrounding the code to be inserted with *{quote:"<html>"}* and *{quote:"</html>"}*. <html><b><i>Warning:</i></b></html> Because text between {quote:"<html>" and "</html>"} is interpreted as HTML, _even inside verbatim paragraphs_, if you need to use the string "{quote:<}html>" in sample code, you must wiki it as "{quote:{quote:<html>}}", or your sample code will be parsed and rendered instead of displayed as written. This does not affect </html>, nor other HTML or XML markup, just "{quote:<html>}". **Boldface and Italics** Text contained between asterisks is rendered *bold.* If you use two or three asterisks in a row, instead of just one, the bold text is also shown at a larger point size. Text between underscores is rendered as an _italics._ All font style markers must start at the beginning of a word (a contiguous text string) and must finish at the end of a word within the same paragraph. You cannot use both at once. However, you can acheive the desired effect like so:{linebreak} {quote:<html><b><i>Italicized bold text</i></b></html>}{linebreak} but at the cost of a linebreak inserted before the text marked up this way. **Lists** If a line begins with the characters "*{quote:*:}*" followed by a space or tab, then that line becomes an item in a bullet list. Similarly, if the line begins with "*N:*" (where N is any number including a multi-digit number, or 0) then the line becomes an item in an enumeration list. Enumeration items are automatically renumbered so the values of N do not need to be in ascending order. Lines that begin with "*{quote:_:}*" are indented like a bullet list but do not display the bullet. You can use this to indent without getting monospaced fonts and other behaviors of <pre>. Make nested lists by adding colons. For example, to make a second level bullet, begin the line with "*{quote:*::}*". **Hyperlinks** Links to other pages are created automatically whenever the *name of another wiki page* is mentioned in the text. (See WikiPageNames.) If you want to put a CamelCase word in your text but you do not want it to become a hyperlink, enclose the name in "{quote: {quote: ...}}". This auto-linking does not happen inside indented verbatim paragraphs. Full *URLs* in the text, such as links to external sites, also automatically become links, but see below for special handling of image links. Such auto-links are created for any URL beginning with "http:", "https:", "ftp:", or "mailto:", without having to use the {quote:{link:}} functionality documented below. *Note:* If the URL ends with ".jpg", ".jpeg", ".gif", or ".png" then the image that the URL points to is displayed inline on the wiki page. You can also create an inline image using markup like this: "{quote: {image: URL}}". Using the {quote: {image:...}} markup allows the image URL to be relative. This allows an image stored in an attachment to be displayed inline. Text of the form "*#NNN*" where the "NNN" is a valid ticket number becomes a hyperlink to the ticket. Text of the form "*[NNN]*" where "NNN" is a valid check-in or milestone number becomes a hyperlink to that check-in or milestone. These hyperlinks only work if the user has permission to read tickets, check-ins, and/or milestones. To create a hyperlink on arbitrary text, use "*{quote: {link: ...}}*" markup. Any text of the form: "{quote: {link: URL PHRASE}}" displays PHRASE as hyperlink to URL. Example:{linebreak} {quote: {link: http://www.google.com Google}} = {link: http://www.google.com Google}{linebreak} The URL can be an absolute URL beginning with a prefix like "http:", or it can be a relative URL referring to another page within the same CVSTrac server. For example, to create a link to a ticket report, one might write: "{quote: {link: rptview?rn=1 Active Tickets}}". <html><b><i>Warning:</i></b></html> The {quote:{link:}} mechanism cannot handle long links. If the entire {quote:{link:}} command will not fit onto one line in the wiki editing text box, the link will get mangled. If this happens, use http://tinyurl.com to create a shorter, redirect URL. **Horizontal Lines** A horizontal line (the <hr> markup of HTML) is generated whenever four or more "-" or "=" characters appear on a line by themselves:{linebreak} *{quote: ----}* **Other Markup Rules** The special markup "*{quote: {linebreak}}*" will be rendered as a line break or hard return. The content of "*{quote: {quote: ...}}*" markup is shown verbatim, including wiki commands, e.g.:{linebreak} {quote: {quote: {quote: ...}}} ---- edit
509 doc active 2005 Nov cpb 2005 Nov cpb 3 3 need online help/user manual There's a lot of features in CVSTrac that people don't know about at first glance because, well, there's no documentation about actually _using_ it. edit
513 new active 2005 Nov anonymous 2005 Nov cpb 3 3 svntrac should support more than one svn repository To be functional comparable with cvstrac, svntrac should support somthing like a parent_svn_dir and include all svn repos in this directory. At the moment we can only use one repo, which is the same like one cvs module. edit
532 new active 2005 Dec chorlya 2005 Dec chorlya 5 5 Show text length when ticket summary too long When one-line summary is too long it would be nice to display how long that text is at that time. This is just to give user an idea of how far off s/he is. edit
265 new active 2003 Nov anonymous 2005 Dec drh 2 3 Accessibility support Can you add accessibility support to forms, please? Just add additional atributes to buttons and input fields, example: <input type="text" name="t" size=70 value="" accesskey="T" title="The title of the ticket, press Alt-T to switch to this field"> ... <input type="submit" value="Submit" accesskey="S" title="Press Alt-S to submit new ticket"> edit
549 new active 2006 Jan anonymous 2006 Jan drh 4 4 Wish: trigger on Wiki changes (e.g. for mail) My team would really appreciate a mail notification on wiki changes. Also, the permission system is great; but still, every Wiki webmaster needs to take measures against illegal content and vandalism. I understand that you would not hard-code mail in here; but could we duplicate or reuse the mechanism that is used for ticket changes? -Karsten edit
557 new active 2006 Jan cpb 2006 Jan cpb 5 3 create new tickets by e-mail It'd be useful to be able to create a ticket around an e-mail message. I get a lot a bugs reported to me by mail and cutting and pasting them and any associated attachments is a pain in the butt. edit
563 doc active 2006 Feb anonymous 2006 Feb drh 1 2 Target fix version When creating a ticket, i would like to be able to set a string that gives the version the bug need to be fixed for (eg: 1.1) Would be even better if this could be a drop down and that we could set some marketing version somewhere in the setup edit
572 new active 2006 Feb chorlya 2006 Feb chorlya 4 4 Few usability improvments to HTML markup It would be nice to tidy up our forms a bit. I'm thinking of adding taborder to few forms and adding =<label>= to every checkbox/radio button. edit
573 new active 2006 Feb cpb 2006 Feb cpb 5 4 should flag links/pages as NOFOLLOW and/or NOINDEX If you do a search on cvstrac.org, you notice that wiki history seems to be getting index. It wouldn't be a bad idea to have NOINDEX, NOFOLLOW on the meta sections of certain pages (although we don't really control that part of the header) as well as a rel="nofollow" on internal links. rel="nofollow" on all external href's wouldn't hurt, either. edit
577 new active 2006 Feb cpb 2006 Feb cpb 4 3 need better wiki spam protection See WikiSpam for discussion of evils. Or just watch the /timeline. edit
578 new active 2006 Feb cpb 2006 Feb cpb 5 4 search should allow "free-form" queries It would be _really_ nice to be able to perform quick report-like queries without having to go through the hassle of creating an actual report. The ideal approach would be to allow the /search to take a SQLite query and run it through the report filters, then generate a decent generic output table with some results. Linkable results would be nice, but no need to get picky. Obviously only users with query permissions would be able to use this feature. edit
581 new active 2006 Mar cpb 2006 Mar cpb 5 3 better control over report field formatting One thing that's been holding me back from submitting #226 is the insufficient control over report formatting. For example, I can't create a "recently editted wiki pages" where each page name is a link to the wiki page. Ditto for checkins. Unless I go with the underscore column name, which really makes a mess of layout. Only tickets have the ability to be links, but they're a bit weak since, well, they bring in the _edit_ link and you don't get any context info (link titles, etc). So we need a way to provide better formatting info for reports without breaking existing stuff. edit
591 new active 2006 Mar anonymous 2006 Mar drh 3 3 CVSTrac should support HTML headings HTML supports markup for headings (<h1>, <h2>, etc.), which basically anyone uses for structuring of non-trivial documents. CVSTrac does not appear to provide for a way to generate such headings. The most common WIKI syntax for headings I can find is that any paragraph which ends and starts with equal signs. Examples: = Hello = (will create <h1>Hello</h1>) == World = (will create <h2>Hello</h2>) == World === (same as above) edit
592 new active 2006 Mar chorlya 2006 Mar chorlya 5 4 Nicer error pages Once, if ever, this ticket is closed we should have: *: Special error page for schema upgrades. Since that can take some time it would be nice to let users know that some maintenance is in progress and that they should try again soon. *: Error pages that use user defined header and footer. *: Nicer, generic, error pages if above fails due to db locking Implementation details are not known at this time. edit
596 new active 2006 Apr cpb 2006 Apr cpb 5 4 file repository markup could use anchors Now that it's possible to reference files as wiki markup, and the /fileview page formats the output as HTML, it's feasible to allow for markup like: {quote:/path/to/file.c#line} or {quote: /path/to/file.c#function}, where the /fileview generator creates HTML anchors for each line/function. Of course, line numbers are easy while function references are language specific. But line numbers break with revision changes. Still, worth thinking about. edit
598 new active 2006 Apr anonymous 2006 Apr   5 2 Add pagination display function to browsing tickets I create many tickets, i want to browse these tickets on pagination display mode. Can you add pagination display function to browsing tickets? edit
600 new active 2006 Apr chorlya 2006 Apr chorlya 3 4 Wrap common HTML form elements in helper functions Since we're adding more markup to forms (label, fieldset, tabindex and accesskey) its getting really ugly with all that markup in code. We should wrap most of the form elements in helper functions and even wrap those in higher level helpers. Such as =cgi_radio_fieldset()=. edit
289 new active 2004 Feb chorlya 2006 Apr   5 5 duplicate tickes Hello, First of all, thank you for an excellent product !. It would be nice if cvstrac could mark a ticket as a duplicate of another ticket (as is in bugzilla). Something similair can be done with the "Derived from" option when editing a ticket, but a "Duplicate from" would be more clear. With regards, Robbin edit
604 new active 2006 Apr cpb 2006 Apr cpb 4 4 embedded scripting language support Enhancements like #183 and #405 are dependent on external programs. This makes a CVSTrac install a bit more complicated to manage. It would be nice if, instead of calling external programs, we could just invoke subroutines in an embedded scripting language. I.e. rather than "program %S1 %S2 %S3", we'd call a subroutine with the list of substitutions as parameters: sub program { my %parms = @_; print $parms{S1}; } Example is in perl, but any embeddable language with associated arrays (Tcl, Python, Ruby?) would work. It would be necessary to make CVSTrac support more than one language, obviously, because there's no chance that we'd ever get concensus on a single language. Handling of script output is a bit more complicated. Would probably require exporting =cgi_printf()= and/or =cgi_append_content()= or something... edit
608 new active 2006 May cpb 2006 May cpb 4 4 need to separate the capability to see source and check-ins In situations where you don't want to publish _source code_, you may (usually?) still want to allow users the ability to see the text of check-in messages, times, branches, etc. In effect, this is no different from published a full CVS ChangeLog with binaries. Since breaking existing configurations is a bad thing, the best way to handle this is to add a new capability, like "v" for "okChngView". edit
622 new active 2006 May anonymous 2006 May drh 3 3 proper base URL handling when running behind mod_proxy I run cvstrac on a different server than the one actually fielding the request, using Apache mod_proxy to do reverse proxying. This is a fairly common configuration, specially when you are trying to make cvstrac appear as a logical part of a larger site. When serving up a RSS feed, however, cvstrac embeds absolute URLs to the proxied server's URL rather than the user-visible one. The following patch fixes this in my environment (we use Dan Bernstein's {link: http://cr.yp.to/ucspi-tcp.html UCSPI-TCP} rather than inetd to manage connections). It does make the assumption that mod_proxy's =X-Forwarded-Host:= header (the one sent by the browser) will come after the =Host:= header generated by mod_proxy. It is quite likely cvstrac configured to use CGI will need equivalent patching elsewhere to get the same information from an environment variable. <html><pre> *** cgi.c~ Fri May 26 07:20:28 2006 --- cgi.c Wed May 31 16:09:36 2006 *************** *** 2507,2512 **** --- 2507,2514 ---- putenv(mprintf("HTTP_REFERER=%s", zVal)); }else if( strcmp(zFieldName,"host:")==0 ){ putenv(mprintf("HTTP_HOST=%s", zVal)); + }else if( strcmp(zFieldName,"x-forwarded-host:")==0 ){ + putenv(mprintf("HTTP_HOST=%s", zVal)); }else if( strcmp(zFieldName,"content-type:")==0 ){ putenv(mprintf("CONTENT_TYPE=%s", zVal)); }else if( strcmp(zFieldName,"cookie:")==0 ){ </pre></html> edit
628 new active 2006 Jun cpb 2006 Jun cpb 4 3 should be able to feed RSS incrementally One thing that's annoying is that when an RSS reader requests 600 days of data, we _send it_ 600 days of data. Even when it last requested an update five minutes ago. It would be nice (for both the client _and_ server) to use something like =If-Modified-Since= to limit the length of the search to just what has changed since the last update. edit
629 new active 2006 Jun cpb 2006 Jun cpb 5 4 should be able to use ticket view elements as anchors RSS and Timeline reference specific updates... It'd be nice to have links in either go to an anchor within the ticket. Appended remarks might be a bit tough since converting the existing ---- syntax to anchors is iffy and editting the text itself may destroy things.. edit
639 new active 2006 Jul anonymous 2006 Jul cpb 5 4 Suggestion: grouping and ordering for list of available report formats Since the list of reports on our system grows larger every day, it would be nice to have a means of grouping reports in a tree like structure. For the time being it would suffice to let the group structure to be only one level deep. Example: <html> <ul> <li><b>Development</b></li> <ol> <li><a href="rptview?rn=14">Active tickets by priority</a></li> <li>...</li> </ol> <li><b>Quality assurance</b></li> <ol> <li><a href="rptview?rn=63">Showstoppers</a></li> <li><a href="rptview?rn=7">Ticket counts by status</a></li> <li>...</li> </ol> <li><b>Misc.</b></li> <ol> <li><a href="rptview?rn=70">Wiki Search</a></li> <li>...</li> </ol> </ul> </html> One-level grouping could be done by one new field named 'group' in REPORTFMT, which could be edited on the rptnew and rptedit page by two input fields: "Assign to report group" (drop list showing previously defined groups) and "Assign to new report group" (free edit field). For report ordering I don't have a simple suggestion other than providing a new field named 'ordinal', which the user can edit freely. edit
641 new active 2006 Jul cpb 2006 Jul cpb 5 4 query-based user defined fields It's possible to define custom fields with a list of possible values. It'd be even better if we could define a custom field which pulled the list of values/descriptions from the database, similar to how the *option()* function works. This would allow, in particular, for milestone-based custom fields (fixed in, target, etc). edit
646 new active 2006 Aug anonymous 2006 Aug cpb 5 4 more hyperlinks in reports Output from reports already hyperlink to ticket numbers --- is there some way hyperlinks could be added to the output from reports for other resources as well? For example, it be great if I could actually click on the Wiki pages or developer names in {link: rptview?rn=70 this report}. Maybe there is already a way to do this, but it eludes me (cvstrac newbie) edit
660 event active 2006 Sep anonymous 2006 Sep drh 1 2 How to convert "cvstrac" database to "mantis"? I've been a happy cvstrac user for quite a while. But I've had to change my server, and I cannot run cvstrac on it any longer (due to IT constraints). I have set up "Mantis" (http://www.mantisbt.org/) as a reasonable substitute, but I have an existing cvstrac database I would very much like to convert to "Mantis" so I don't lose any of my data. My first attempt at conversion was not very successful. Does anyone have any pointers or (hopefully) a script to convert the format? Thank you. edit
661 new active 2006 Sep anonymous 2006 Sep cpb 4 3 More customization for timeline Say, *:show the given person's activity; *:show given subsystem's activity; *:show given file's activity;{linebreak} etc. edit
662 new active 2006 Sep anonymous 2006 Sep cpb 5 5 That would be nice if wiki can output offline version itself Document group in our project complained about that they had to manually strip the document maintained in the cvstrac's wiki. So our daily build can't included document right now.Maybe this can be done in c source? Or we can perl/python that function outside. edit
671 new active 2006 Nov cpb 2006 Nov cpb 4 4 Should identify branches when they're tagged CVSTrac creates new milestones from =cvs rtag= operations. It would be _really nice_ if it figured out which tags were branches and flagged them somehow. Perhaps this information could go into a separate TAGS table. Otherwise, the only way to find out if a branch is a branch is when someone commits on the branch. This may never happen. Branch detection can't be done by looking at =CVSROOT/history= (the -b option doesn't come through in any way), so it'd require looking at an arbitrary RCS file under the tagged module directory to determine whether it's a branch or not. edit
673 new active 2006 Nov anonymous 2006 Nov cpb 3 3 custom block wiki markups should have additional substitutions I am trying to write a custom block markup that would allow inserting inline graphs produced by graphviz, similar to the excellent {link: http://twiki.org/cgi-bin/view/Plugins/DirectedGraphPlugin dot plugin for Twiki}. The way it would work is the external program would take %b, unescape it and run dot/neato/circo on it (depending on =%m=), attach the resulting PNG to the wiki entry (by accessing the CVStrac table directly) and output an =<img>= tag linking to it. The program would use a MD5 checksum to avoid rerunning graphviz if it had already been run on the exact same text (the attachment would be in effect a cache). To do this, the program needs: *: the name of the SQLite database (=%B= ?) (although that could be hardcoded in the custom markup declaration) *: the name of the wiki entry or ticket (e.g. =%W=) so it knows which entry to attach it to. *: optionally a list of attachments to the wiki page (=%A= ?) These two substitutions would make truly powerful block markups easier. In the meantime, I will leverage the fact SQLite does not enforce referential integrity to attach the generated PNG nowhere. edit
663 new active 2006 Sep cpb 2006 Dec cpb 3 3 check-in diffs (and other things) should have a maxmimum size So a colleague checked in a large number of data files. The resulting check-in diff weighs in at about 50MB. This isn't the first time I've seen this. Given that a "patchset" option is now available, some way to restrict the size of "normal" pages would be nice. edit
677 new active 2007 Jan cpb 2007 Jan chorlya 5 3 SvnTrac should be able to show file/directory properties Subversion is able to store arbitrary properties for objects under revision control. It would be useful if SvnTrac provided a way to discover these properties (see svnlook's proplist/propget commands). edit
678 new active 2007 Jan cpb 2007 Jan cpb 5 4 attachment upload should have a "replace" option When uploading a new version of an attachment (stypesheets, say) it would be nice to toggle on option to replace existing attachments with the same name. This would save a whole bunch of manual deletes. It'd probably have to work by completing the upload, then, after getting a list of other attachments, redirect to /attach_del with a list of atn's to delete. The user would then get redirected back to the original page after deleting (or cancelling). edit
680 new active 2007 Jan cpb 2007 Jan cpb 4 4 should be able to toggle default milestone type The default milestone type is "Release". However, 99% of the time I only manually create "Event" milestones because: 1: I use =cvs rtag= to generate releases 2: on some installs, I only use milestones for keeping short notes It'd be useful if it was possible to toggle the default, ala ticket defaults. It'd also be nice to change the default for the type of milestone created for rtag's, since I always have to change them to "Release". edit
681 new active 2007 Jan cpb 2007 Jan cpb 5 4 report docs very ticket-specific For example, we have instructions like =Enter a complete SQL query statement against the "TICKET" table= and the format hints only output the TICKET table. Reports against other tables are very handy. Might be useful to have something like a report wizard that allows the user to select a table (to see the schema) or something similar. Certainly, the documentation should be a bit more generic. edit
682 new active 2007 Jan cpb 2007 Jan cpb 4 4 Ticket remarks should include HTML anchors This would allow us to reference specific comments in tickets. I suspect that the "best" solution to this sort of thing would be to introduce a {quote:{remark: <user> <when>}} markup which automatically formats everything appropriately and inserts an anchor (a per-page counter would probably be best). This would be extremely useful for RSS links, particularly in RSS readers which can automatically load the web page rather than view the RSS (i.e. akregator). edit
685 new active 2007 Jan cpb 2007 Jan cpb 4 3 spam keyword handling should allow for embedded spaces In other words, the *search()* function should allow for things like (double) quoted text or something similar. edit
575 new active 2006 Feb cpb 2007 Jan cpb 3 3 attachments need to be changed to BLOBs SQLite 3 directly supports BLOBs, so we can eliminate the encoding/decoding magic currently used for attachments. This will require a schema upgrade. edit
368 new active 2005 Jan anonymous 2007 Feb drh 2 2 Enable to set output encoding or use utf-8 by default I'm using cvstract for acces to the repository with various language mutations of processed texts. It is boring to set manualy utf-8 encoding in my browser for each page with source diff i want to review. It would be nice to have configuration option for output encoding edit
686 event active 2007 Mar anonymous 2007 Mar cpb 5 4 Wiki limit Trying to add a question to FAQ Wiki page, returns an error. Question not added. Edit: "Frequently Asked Questions" Too many external links for one edit! If i write queston again deleting all text except my question (in case it works like an automatic reply) gets the human confirmation page but question is not added. I don't know if it has been added or not. edit
688 todo active 2007 Apr anonymous 2007 Apr cpb 3 3 Login verification in IE 7 Login verification seams not to work with standard setup of Internet Explorer 7 edit
692 new active 2007 Apr cpb 2007 Apr cpb 5 4 Wiki diff view should also be able to show page history list It would be nice, when perusing a long list of wiki history, to only see diffs. Unfortunately, the page history sidebar disappears when in diff mode. edit
691 new active 2007 Apr chorlya 2007 May chorlya 4 4 Show version info in cmdline output When printing help/info messages version info should be shown as well so you can easily know which version your binary is. edit
653 new active 2006 Aug anonymous 2007 May cpb 4 3 Persistent login Users are required to re-login for each session. It would be nice if there was an option to automatically login without having to explicitly click the login button every time. This is especially annoying as it is easy to forget to login before editing a wiki and having your edits flagged as anonymous. edit
690 new active 2007 Apr chorlya 2007 May chorlya 4 4 Add option to hide deleted files in Browse When browsing repository via Browse (short and long view), there should be a link in action bar to show/hide files that are deleted in repository. This should default to "show all" and choice should probably be preserved via cookies. edit
516 new active 2005 Nov anonymous 2007 May drh 2 4 Use PAM for authentication Would it be possible to have cvstrac authenticate users by using PAM? edit
693 new active 2007 May anonymous 2007 May cpb 4 4 How to add more SCM I want to add monotone scm to cvstrac system. Should i base it on current cvs, svn and git code or start from begining? Have cvstrac standard or default commands that are converted to cvs, svn, git commands or each scm has it's own commands? Any howto about add other scms? Thanks edit
695 doc active 2007 Jun cpb 2007 Jun cpb 5 4 HOWTO migrate to GitTrac Notes on converting from CVS/CVSTrac to GIT/GitTrac are at GitTracMigration. This ticket is just meant to accumulate comments and question without cluttering up wiki pages with discussion. edit
701 new active 2007 Jun cpb 2007 Jun cpb 5 4 External tools should allow just URLs External tools allow admins to add new things to specific action menus. It would also be really handy if we could just add links to other services. webcvs, Subversion's webdav browser, gitweb, etc. CVSTrac is generally good, but more specialized apps are _often_ better at their specializations. This could be done just by calling =is_url()= on the command. We'd still do all the substitutions, then embed into the action menu as-is. edit
706 new active 2007 Jul anonymous 2007 Aug cpb 4 3 new ticket defaults should support ALL fields (e.g. default descriptio some organizations break up description into sections or otherwise have a template/guide for users to follow (since they often forget to include important bug details, etc.) cvs has checkin comment templates that are similar in concept. The user is free to delete the text, but they have to "want to". Why not just be able to add default text for any field just like the ones that are setup-able. This saves time in a process by having fewer tickets get rejected or handed back to the submitter for revision or clarification. Another nicety would be to set "required text" just like cvs checkins can have. Then the submitter would immediately know they didn't do enough boiler-plate stuff . . . or that it was impermissible to delete the template. edit
253 new active 2003 Sep anonymous 2007 Aug drh 1 1 Changing commit logs in the repository cvs admin -mrev:msg - Replace the log message of revision rev with msg. It would be nice to have the option of having the changes to commit logs in cvstrac propagate into the repository, using this command. edit
495 new active 2005 Oct cpb 2007 Aug cpb 5 3 need some benchmarking/tracing support We're finding the need for some benchmarking and tracing logic. #494, for example. It's possible to benchmark individual queries with some work, but repeatability is poor. What I'm thinking of doing is creating a DummyTrac SCM module which allows for generation of a single common benchmarking database and enables benchmarking support in various other places (i.e. at the db_query level). Comments? edit
300 new active 2004 Mar anonymous 2007 Aug drh 2 3 Enable web server authentication We are using cvstrac with many other application on a group portal. The portal is responsible for authenticaton and authorization. With the current setup, users have to "double-login" to access the cvstrac, under their own identity. It will help if cvstrac will include the following "Web Authentication" modes (using the REMOTE_USER environment variable). No password will be required for users: *: Everyone: accept any value of REMOTE_USER. If the user does not exists, automatically create the user entry. *: Selective: the REMOTE_USER must match an existing entry in users. edit
716 new active 2007 Aug anonymous 2007 Aug cpb 4 3 Simplify file browser icons HTML code The file browser uses code: <a href="rlog?f=%T(zName)%T(az[j])"><span class="icon file"></span></a> <a href="rlog?f=%T(zName)%T(az[j])">somefile</a> .. while it doesn't really make sense when using CSS, I'm submitting the patch that will rather make: <a href="rlog?f=%T(zName)%T(az[j])"><span class="icon file"></span>somefile</a> .. and save some HTML. This is litterally the same.. plus we may add some padding to =.icon= add CSS to achieve same result. The patch also adds =.report= class to the ticket report table, so it can be customized in CSS. Best regards. edit
718 todo active 2007 Sep anonymous 2007 Sep anonymous 3 3 Assistance with upgrade from 1.1.5 to 2.0.1 Morning, I am preparing to upgrade a cvstrac instance from 1.1.5 to 2.0.1 and im uncertain what the best path to success is. Ive got a testing instance going so i'm going to go and probably break it a few times but in the meantime I figured it wouldn't hurt to seek your input. Ive read the two docs at: http://www.cvstrac.org/cvstrac/wiki?p=DownloadCvstrac http://www.cvstrac.org/cvstrac/wiki?p=CvstracUpgrading I was attempting to compile the attachdump.c prog and failing miserably. Is this tool actually required? [root@crbt-cvstrac-test cvstrac-2.0.1]# gcc -g -O2 -o attachdump /usr/src/cvstrac-2.0.1/attachdump.c /tmp/ccVWOLh2.o(.text+0x288): In function `main': /usr/src/cvstrac-2.0.1/attachdump.c:127: undefined reference to `sqlite_open' /tmp/ccVWOLh2.o(.text+0x2b6):/usr/src/cvstrac-2.0.1/attachdump.c:133: undefined reference to `sqlite_exec' /tmp/ccVWOLh2.o(.text+0x2d2):/usr/src/cvstrac-2.0.1/attachdump.c:138: undefined reference to `sqlite_close' collect2: ld returned 1 exit status Any suggestions you can offer would be appreciated. edit
231 todo active 2003 Jul anonymous 2007 Oct cpb 4 1 Replace all hard-coded colors with CSS classes Customizing CVStrac can be a hellish nuisance with all the hard-coded colors within the C code. This is especially true in report tables, where there is no way to customize the rows themselves. My proposal is to make CVStrac layout and coloring fully CSS-based. In this fashion, a report table would have the following tagging: <table class="reportTable" ...> <tr class="reportHeader"><td class="reportHeader ticketNumber">... <tr class="reportRow activeTicket">...<td class="activeTicket ticketSubsys"> ...etc. Note the use of multiple classes to enable full customization. You can, for instance, say: .reportTable { font-size: 8pt; } .reportHeader { font-weight: bold; text-align: center; } .reportRow { padding: 2px; } .ticketNumber { font-face: Courier; } .ticketSubsys { font-style: italic; } .activeTicket { background-color: orange; } This would make the entire table 8pt size, the report headers bold and centered, give each report row extra padding, and give distinct coloring and fonts to both active ticket rows and specific columns. I might find the time to make a patch, but I'd like to have some feedback first as to whether this would be incorporated. edit
727 new active 2007 Oct anonymous 2007 Oct cpb 4 3 Adding Mercurial (hg) support Is it possible to add Mercurial (hg) support? I'm thinking on something called hgtrac working like gittrac (as both Git and Mercurial are distributed). edit
730 new active 2007 Nov anonymous 2007 Nov cpb 4 3 Need Massive Edit Mode (Multiple Items in one view) We are using svntrac quite regularly and one of the most frustrating things right now is that there is no easy way to do massive editing of tickets. For instance, as we are nearing a release milestone, we want to review the tickets and do a quick in/out analysis and there is no way to do things like setting the priority, or changing other fields all from one page. For example, suppose you have a field called "Target Release" and your normal priority columns and you are looking at 200 tickets that you own. Often from the description you could figure out if this is for this release or the next release and also set the appropriate pulldown. The resulting report would have pulldown items instead of text for some of the columns and when you push update, it would update all of them at once. edit
735 new active 2007 Dec anonymous 2007 Dec   4 3 add some flexibility in CGI args With some simple web servers (SHTTPD, Abyss) it is convenient to pass the script pathname as an argument to svntrac.exe. For example, you could set up Abyss to use svntrac.exe as a script interpreter with args "cgi <dir> %1" where <dir> is a pathname, and the %1 is filled in with the path to the cgi script. The name of the script is the name of the project -- the script file can be empty since svntrac doesn't open it; it can even be the database file. But svntrac doesn't strip the directories from the pathname so supplied as it normally does with the project name obtained form the URL or HTTP_POST. So, I added this feature in main: if( argc>=4 ){ /* The project name is specified on the command-line */ #ifdef CVSTRAC_WINDOWS /* e hack for Windows CGI */ i = strlen(argv[3]); while( i>0 && argv[3][i-1]!='\\' ){ i--; } g.zName = mprintf("%s", &argv[3][i]); #else g.zName = argv[3]; #endif cmdlineProj = 1; where the new stuff is in the {quote: #ifdef CVSTRAC_WINDOWS} section, but other than the path separator, is not really Windows specific. edit
737 new active 2007 Dec anonymous 2007 Dec cpb 4 3 Search should allow phrase searches Search module should allow searching for phrases, when the query is inside double quotation marks. Example: Searching for "sqlite database" in tickets should only return tickets with the fixed text "sqlite database" inside title, description or remarks. edit
564 new active 2006 Feb anonymous 2008 Jan   5 4 More i18n I have written more i18n patch. *: Multi-language message support using gettext *: Support content negotiation (need -DCVSTRAC_CONTENT_NEGOTIATION) *: Accept 8bit Wiki title (need -DCVSTRAC_8BIT_WIKINAME) *: Add the tag of format {wiki: TITLE TEXT} (need -DCVSTRAC_8BIT_WIKINAME) To write message catalog, $ make cvstrac.pot $ cp cvstrac.pot <lang>.po $ vi <lang>.po # make translation and add <lang> to TRANSLATED_LANGS of main.mk. edit
187 new active 2003 Mar anonymous 2008 Jan drh 4 2 cvs import/commit "Browse Repository" annoyance I've created some new repositories to manage our projects and i first realized why the files aren't listed in the "browse repository" section. This seems to be described also in #88 and #113. I ever use cvs import to checkin new source trees. That's why cvstrac didn't update the tables (if this is right). The problem is, using "cvs add" takes time, can only be used with a "find . -print -exec cvs add {} \;" to automate many file checkins, and "cvs add" is even not mentioned to use for cvs with initial checkins of complete source trees. There are two ideas how to "fix" this problem 1. Name the "Browse - Browse the CVS repository tree." menu as "Browse - Browse changed files in the repository tree.", but this is not the way i want to work with cvstrac *2.* Find a way to "reconstruct" the filechng (and other influenced tables) as setup-menu or as "realtime-sync-filetree-on-browse" which seems to be useful... Is there another way to fix this? cvs rtag tagname modname doesn't work, did i used the rtag command wrong ?` Bye, Bjoern edit
741 new active 2008 Feb anonymous 2008 Feb cpb 4 3 Allow milestone wiki documentation We use milestones to mark planned releases of our developments and it would be usefull to have a separate simple text field for the milestone name and another wiki area where we can enter a description of the milestone purpose. edit
35 new active 2002 Jun anonymous 2008 Mar drh 4 4 Groups of users, assigning bugs to groups We use CVSTrac at work. One of the users submitted the following feature request to me. I'm forwarding it to you. Here goes: When creating tickets, we are restricted to assign it to only one person. A lot of times I'd like to assign it to more than one person, or to just "Java programmers" in general, or something like that. It seems like having groups of programmers would be a powerful feature, and we could definitely use it in our organization. edit
42 doc active 2002 Jun anonymous 2008 Mar drh 3 3 add description of ticket types/et cetera Please, add to the documentation (or include as part of cvstrac binary) explaintations of ticket types (for example what is "action item"), and other similar things. Also, if you have time, add some general recomendation from your own use of the system, some helpful techniques. thank you. edit
745 new active 2008 Mar anonymous 2008 Mar cpb 5 4 Putting CVSTrac timelines on project websites A project timeline such as appears at http://www.ceamus.com or http://burnbox.lazarusid.com is nice to have. I have implemented it for these two projects using a shell script, an XSL transform, and server side includes. In my project home page I include the following: <!--#include file="timeline.html" --> I run the attached getrss.sh script via cron to provide regular updates (currently hourly, but adjust as appropriate for your project). Because there is a flaw in the RSS generated by CVSTrac, I also run it through fixrss.c (attached) edit
746 new active 2008 Mar anonymous 2008 Mar anonymous 4 3 Wiki timeout causes data loss I'm not the cvstrac administrator at my site, but I do use the tool heavily. I FREQUENTLY run into situations where I'm editing a wiki entry, but while I'm doing so, my login cookie expires, and I get prompted for a new login. *poof* all my editing is lost. Unlike twiki (which I once used), you can't "go back" to a page you were editing, because those pages are created by "POST" requests, and are therefore not cached (I surmise). You also can't "checkpoint" pages as you're editing. Is there any way to have the tool either (a) capture the data provisionally when a timeout is detected, or (b) not use POST to edit pages, so that I can at least go "back" in the browser to cut my edits and paste them into a new session. In a related problem, I sometimes run across the magic key sequence that closes tabs, and, again, all my updates are lost in a flash, since I can't "go back" and get to my edits again. edit
748 new active 2008 Apr anonymous 2008 Apr cpb 4 3 Central Management is missing I am deploying this as a central project management interface at Frazer-Nash. Myself and another programmer are allocated to doing this. We also have a graphic designer (complete with high end Mac :-) who is allocated time to the project. We would like a central user and project management interface to handle multiple projects. I have knocked up a cgi script in perl to handle add/list/deleting of all the projects This is a dump of the (simple) interface (with the names changed to protect the guilty :-) ---- <HTML> <CENTER><H1>CVSTrac Project Admin</H1></CENTER> <P ALIGN="right"><A HREF="link">How to use this</A></P> <CENTER> <H2>Current Projects:</H2> <TABLE BORDER="0" CELLPADDING="0" CELLSPACING="0"> <TR><TD>[radio button]<A HREF="http://server:8001/One/">One</A></TD></TR> <TR><TD>[radio button]<A HREF="http://server:8001/Two/">Two</A></TD></TR> <TR><TD>[radio button]<A HREF="http://server:8001/Three/">Three</A></TD></TR> </TABLE> <BR> [Delete Selected Project] (<submit)<BR> <BR> <HR> <BR> Name: [text input][Add New Project] (<submit)<BR> <BR> <HR> </CENTER> <A HREF="link_to_home">&lt;Wiki Home</A> </HTML> ---- This page acts as a project add/delete page also you can go to the project via the link. This is obviously pathetically simple, but it means I can get a non-tech to manage this and let me get on with the tech stuff. There seems an obvious entry point to add just this in the / place which currently does a "page not found". Do you have any plans for doing something like this? ---- Also we would like to assign users to projects via a central interface . The problem is that users currently have differently set up accounts (particularly the username and password) on different projects, so they as coming back to me to try and fix this up. Any plans on this too? We have looked at engineering this ourselves and will be looking at it, but it seems either a large change to the C source or some pretty hacky SQLite stuff in perl/php. Please email if you need more because I would really like to get this working well for us. edit
749 new active 2008 Apr cpb 2008 Apr cpb 4 3 Inspection enhancements Lately I find we've been making extensive use of inspections as a general form of checkin comments. I think a few changes might be useful: First, the existing inspection is a text "result" field. To make that more useful, it might be nice to make it available as a drop-down list of options similar to how the "extra" ticket fields work. That is, could be free form, could be a strict list of "pass", "fail", "tested", "rejected", etc. Second, a wiki comment field is needed for inspections. For backwards compatibility, this should be separate from the "result" field. Those are the two main things. There might be more as I think of them. edit
747 event active 2008 Mar anonymous 2008 May cpb 1 1 Does cvstrac.exe work? How (step by step wiki)? The precompiled binary inside cvstrac-2.0.1-MinGW-20071107.7z just crashs (unhandled exception) with the command: cvstrac.exe cgi Does someone run cvstrac.exe as a CGI Apache (Windows build) program? edit
754 todo active 2008 May anonymous 2008 May cpb 4 3 can cvstrac be setup to use with existing cvs repository? first, my apology for posting question here, i don't see mailing list. my question is, can cvstrac be setup to use with existing cvs repository? how would i go about doing this? would it doing so damage or corrupt the existing repository in anyway? thx. edit
753 new active 2008 Apr anonymous 2008 Jun cpb 2 3 Customizable workflow A simple work flow would be nice to have. Let's set up a state machine with the status and their out-transitions. new review,ignore review active,new ignore new active fixed fixed reopen,verified reopen fixed verified closed The first column is the status in CVSTrac, and the second column is their out-transition that shows the next status. For the first row, =new review,ignore=, it means the user can only set _new_ ticket to _review_ or _ignore_, the user can not directly fix the ticket without review. So, we may just utilize the status customization page to define the state machine. edit
640 new active 2006 Jul anonymous 2008 Jul cpb 5 5 IE can't render the css-patches correctly IE can't render things like '<span class="icon ticket"></span>'. Doesn't support '.icon' or can't recognize things like '\content: "\00002a"',I'm not sure. Maybe try another way draw the icons? edit
771 doc active 2008 Nov anonymous 2008 Nov cpb 4 4 Description of quote in the formatting hints The description for the _quote_ markup says only "display _text_". From my point of view (as a non native english speaker) it seems like this markup is for quoting text and to display it in italic font. It would be good to have an additional comment that says that the part inside the quote will be shown literally without evaluating any other markup in the "quoted" text. edit
772 todo active 2008 Dec anonymous 2008 Dec cpb 3 3 How can we know abt next release Hi,{linebreak} Currently, we are using CVSTrac (ver 2.0) in our organization. How can we know about your next release? Practically, it is highly _impossible_ to check each and every day about new release in the internet. Do you have any other approach, in such a way that we'll get the periodic information about cvstrac new release. edit
778 new active 2009 Jan anonymous 2009 Jan   3 3 Timeline report: list items and/or commit by one or more usernames It would be very useful to see the (very useful) timeline report by user or list of users (e.g. in some delimited manner). A box in the bottom section would be good for this. Couple of years back, i patched the code in an old cvstrac version to do this, but it would be useful if this was supported in the main release. thx edit
780 new active 2009 Jan anonymous 2009 Jan cpb 4 4 need "Module prefix" equivalent in svntrac i'm using cvstrac and now also svntrac but need to restrict the svn repository access to exactly one subtree - the same is implemented in cvstrac as parameter "Module prefix" - this parameter is missing in svntrac. edit
779 new active 2009 Jan anonymous 2009 Jan   4 3 Backup/restore : allow to define the destination file via fulle path This will be very useful as e.g. a backup may be required to be stored to a different file system than the one where the main file is. using a symlink for .bu file does not work (symlink is removed by the backup operation). edit
782 new active 2009 Jan anonymous 2009 Feb cpb 5 4 ability to modify status of ticket when appending a remark It would be nice that when you are appending a remark to a ticket, you have the ability to edit the status of the ticket at the same time. It would save the step of appending a remark and then going back and editing the ticket to make a change to the status edit
784 new active 2009 Feb cpb 2009 Feb cpb 4 3 milestones could/should be non-numeric The whole _milestone as special check-in_ approach is restrictive and non-obvious. _Every_ SCM represents tags and branches and whatnot as symbolic strings, so why are we forcing developers to think of them as just numbers? {quote: [release2_0_1]} is far more meaningful than [876], and distinct from a check-in. In GIT, {quote:[heads/master]} or {quote:[tags/1.32]} would be a lot more useful than just a number. Separating check-ins and milestones would also solve #770 nicely. edit
770 new active 2008 Nov anonymous 2009 Feb cpb 3 3 reuse the svn revision number in "chng" table The change number in SVNTrac is separated. It's better to reuse the SVN revision as the change number (=cn=) to avoid the difference between change number and SVN revision number. edit
791 new active 2009 Mar anonymous 2009 Mar cpb 5 4 RSS view of individual ticket history In Firefox, in the location bar, the RSS icon tooltip reads: "Subscribe to *this page*..." Not only did I mistakenly "subscribe" to my timeline a couple of times before I realized it wasn't the ticket history feed, but thinking back about that mishap it seems to me that an RSS version of tkthistory could be relatively easy to implement. As for linking, unless someone has a better suggestion, I'd be enclined to have cvstrac provide the RSS feed URL in a value expanded in the header instead of leaving it in plain text as it currently is. The advantage would be that on ticket pages (tktview, tkthistory at least) instead of linking to the timeline RSS URL, it would provide the ticket-specific history RSS URL. About making an RSS version of tkthistory, I think I might be able to do it. For actually making the URL available in the header properly, I'm not so sure. (I was about to implement it as an external tool in PHP, when I realized it would fit well within your core features. Needless to say it would also make following up on tickets we created at cvstrac.org much easier, though for that, a filtered version of the timeline would probably be even more efficient.) edit
793 new active 2009 Mar anonymous 2009 Mar cpb 4 4 incomplete history file - how to import i have a project which was imported from RCS to CVS two years ago. so the CVSROOT/history file is imcomplete - all changes which were made with RCS before the import to CVS are lost in cvstrac. but "cvs log" shows all the change details from the old RCS sessions. is there a way to "rebuild" CVS's history file to make the complete story import to cvstrac? or any other idea? edit
794 new active 2009 Mar anonymous 2009 Mar cpb 5 4 Replying to remarks This is probably not possible given how cvstrac deals with remarks but it would be nice to have a way to easily include a copy of a previous remark that a person wants to reply to. I know it can be done manually but is it possible for cvstrac to do? edit
795 new active 2009 Apr anonymous 2009 Apr cpb 5 3 Wiki markup for quoting To suit my own use of CVSTrac, I'm about to add a parser for paragraphs made of lines beginning with ">" into blockquote's, much the same way it's already parsing for "_:" into ul's. I don't think that would break existing wiki documents. The way I've seen cpb "fake" quoting for example in #788 I suspect that this minor improvement would be welcome for all. :) (Okay fine, might as well tell you: I wrote a pretty neat e-mail gateway for CVSTrac in PHP, which allows replying as remarks, notifying the relevant contacts, etc. The one problem I have is displaying quoted portions of messages added to description/remarks and this will solve it.) I'm opening a ticket about it to get feedback as to whether I should post the patch here. So, comments are welcome. I'd also like to know if somewhere in this site's wiki it would perhaps be relevant to list {link: http://www.imars.com/software/cvstrac/ my CVSTrac satellite tools page} in case any of them could be of use to others. Maybe ExternalToolsCookbook? edit
797 new active 2009 May ono 2009 May ono 4 4 Download a tarball of source code (currently browsed folder) Subversion has a =svntar= tool, GIT has =git archive= command. Both of them can produce tarball of the source code on fly. The feature of tarball download is present in SVN or GIT browsers for ages, why don't we add it to CVSTrac !? edit
796 new active 2009 Apr anonymous 2009 Jun cpb 3 3 Make the allowed html tags list for wiki markups flexible User defined wiki markups are filtered on output and the tags not in the table of the allowed tags are removed. As I understand this is a security feature, when used for an intranet, this can be a limitation. It would be usefull to set the list of the allowed tags through a setup page in order to adjust to each situation's security requirements. As an example, we would like to create a new markup that generates a <div> of a specific class (to be then used with a common stylesheet applied by an Apache front-end). edit
804 new active 2009 Jun anonymous 2009 Jun cpb 4 3 List the related check-ins of a wiki page If the check-ins' log message includes the wiki page name, the check-ins are related to the wiki page. When we discuss or record some design issues in the wiki page, it is useful to list the related check-ins in the page. This needs the log message to issue the wiki page name when committing. edit
805 event active 2009 Jul anonymous 2009 Jul cpb 1 1 amendment to ticket 801 re. CVSTracNT hang Hi - I'm about to pull out my few remaining hairs... installing CVStracNT on a Windows 2003 server machine that has the CVS repository on it - I created a new DB, I can log in as setup but if I try importing users, creating a ticket, rescanning the repository (i.e. any DB writes it seems) I can see a journal file getting created and the process never returns. I have to stop the service and remove the journal file to continue. Has anyone seen this behaviour before? I've tried everything I can think of (permissioning on the db file, on the service - even run it as administrator) Is there some sort of debugging output that can be enabled so I can see what's going on? Thanks - Dave edit
807 new active 2009 Aug anonymous 2009 Aug cpb 4 4 login re-direct user always to home page, not page where login'd from When viewing a non-home page, if the user chooses to login, they are unintuitively re-directed away from the page that they were viewing, and sent to the home page. So when viewing a ticket or check-in, if I don't login in first, logging them means I have to manually use the browser's back button's history to go back to where I wanted to be at, the original page, albeit now logged in (thus able to comment / inspect, see source). I realize this likely may require minimal {quote: JavaScript}, but it should be trivial. <input type="button" value="angry monkeys" onClick="history.go(-1)"> or <a href="javascript:history.go(-3)">cargo cult edit
310 new active 2004 Apr anonymous 2009 Nov drh 3 3 Provide more "extra" fields All 5 extra fields have been used up for the ACD and SOFTools programs. Please provide at least 5-10 additional extra fields in the CVSTrac system. edit
816 new active 2010 Jan anonymous 2010 Jan cpb 4 3 RFE: Make it easier to edit tickets with curl Use Case: Bulk editing of tickets{linebreak} Desire: An API to gittrac for ticket edits{linebreak} Ideas: 1: sqlite script 1: curl script posting to tktedit 1: API with language bindings Cons: (1) must be kept in sync with ticket_edit(); (2) currently cumbersome (see below); (3) doesn't exist (1), ticket_edit() isn't simple: SELECT MAX(ROWID)+1 FROM tktchng, UPDATE ticket, INSERT INTO tktchng, SELECT MAX(ROWID) from tktchng, UPDATE ticket, DELETE FROM xref, INSERT INTO xref (2), retrieve a login cookie curl -c cookies -d u=setup -d p=setup -d in= http://localhost/login get the csrf token X=encode64(<login cookie value>){linebreak} Y=encode64(<login cookie name>) then post the edit curl -b cookies -d tn=1 -d le=0 -d submit= -d X=Y -d t=NewSubject ... http://localhost/tktedit (2) appears simpler, but the encode64 step and "..." make it complex. The encode64 step is a custom base64 encoding, and "..." means that all fields of the ticket must be sent for the edit to be accepted. (Yes, the encode64 could be avoided by scraping tktedit?tn=#) Two code changes could make (2) as simple as it appears above. First, encode64's zBase could be changed to the more common A-Za-z0-9+/, effectively turning encode64 into base64 (provided by encoders such as openssl base64). Second, ticket_edit can cnt++ when aParm[i].zNew==0, thus allowing for a partial post of ticket fields. Are there problems with the approach in (2)? How important is the cnt input validation in ticket_edit? Is a zBase of 0-9a-z_A-Z~ just for obfuscation? edit
818 new active 2010 Jul anonymous 2010 Aug cpb 5 3 feature request: user interface improvements This is a feature request for user interface improvements. On the ticket lists, there is an edit button at the end. In order to action a ticket, we have to click edit, then perform an action. It would be better to go straight into an action here, providing a window that allows us to add a comment and change the ticket status. This would make the ticketing system easier to use because it is more likely that a user want to action against a ticket, than edit it. If the user wants to view or edit a ticket, he can simply click on the ticket number and then press edit (which he can do already). The action button would basically be the same as the "append remarks" facility that we have now, but the "append remarks" facility would have the additional option to change the ticket status at the same time. (This would be a useful change anyway), because in closing a ticket, typically we want to add a comment. Currently to close a ticket, we add remarks, then edit the ticket, again to change the status. With this proposed revision, we could add remarks and change ticket statuses easily and quickly. edit
822 new active 2010 Nov anonymous 2010 Nov cpb 4 4 adding option to Search, so as to allow results to be ordered by date The CVSTrac-wide =Search= could be infinitely more useful if it was able to sort the results by date (i.e. reverse chronological order typically, newest first) or other orderings (newest/oldest ticket #). While I realize that it is likely doable with Reports, it seems like it could easily be an out-of-the-box option without much work. (=ORDER BY=) edit
826 new active 2011 Apr anonymous 2011 Apr cpb 4 3 How about the user upload a huge attachment file? I found all the attachment files are stored as bob block in sqlite3 database. Yes, tiny files will be ok, but how about large files? So, i want to store attachments as files with record information in the database. However, im not good at programming, so I couldn't do it by myself. Any ideas? edit
828 new active 2011 Jul cpb 2011 Jul cpb 4 3 should be able to replace/update attachments Attachments are useful, but there's no easy way to replace/update them. Presently, deleting an attachment is _hard_. [594] expanded the rules a bit to mostly prevent abuse and allow users to fix their screw ups. Philosophically, I think there's a strong argument that if an attachment is worth adding to a CVSTrac artifact, it's worth keeping similar to a SCM artifact. That is, deletion _should_ be hard, but there should be some concept of revisioning. Because attachments are artifacts outside the SCM, we'd want to be flexible in terms of naming. Filename-based versions would obviously be common. I'm thinking of a dialog selector of "previous" attachments which can be selected to indicate that a new upload supersedes an existing attachment. That can be easily handled by a new "prevatn" field in the database schema. edit
829 new active 2012 Feb anonymous 2012 Feb cpb 5 4 Wiki Page for Inspections needs some tweaking The explanation about inspections on CvstracInspection should be revised: With the additional comment field the limit is no longer 40 characters. And an advice on what to put into which field would be nice. As well, as an option to _edit_ the inspections. This way you could specify the testing that should be done and after some time when the test has concluded summarize it in the short result field. There is also a small typo: "If they check-in is associated" should either be _the_ check-in or _their_ check-in. This is a "Help" page, so it can't just be edited and corrected as any other Wiki page. edit
504 new active 2005 Nov anonymous 2014 Feb drh 2 2 Timeline is not saving date in Defaults - VERY annoying! The Timeline feature is interesting, but rendered only half-useful, because it defaults to "today and backward", no matter what the user wants. Entering a future date (e.g. 2020-Jan-01) so that all milestone dates are visible (highly desirable behavior on my entire development team) and clicking "Make Default" does not save the preferred date. I understand that "Make Default" could end up saving today's date and making the feature of dubious use for those who (if there are any) don't want to see future dates in the timeline, so I would propose a checkbox next to "Make Default" that says "include date in saved default", checked "off" by default. I think this single change would improve CVSTrac's overall usefulness by about 20%. Seriously. edit
836 new active 2014 Apr anonymous 2014 May cpb 4 3 Pass cgi parameters from wiki page into embedded reports I've modified format.c and view.c:embed_view() to allow wiki syntax like: {quote: {report:rn?Param=Blah Title} } So that reports using the cgi() function can receive params from the wiki. This was a quick hack, but it works well enough for my office. The original syntax of {quote: report: rn}} and {quote: {report:rn title} } still work as expected. I'm not sure how to contact the maintainer of cvstrac though. How do I go about submitting a patch? Possibly related to #561 and #742 edit
837 doc active 2015 Mar anonymous 2015 Mar cpb 4 3 how to install cvstract on the openBSD Hi, how are you. I am the new person for the CVSTrac. I tried to install cvstraac on OpenBSD, but I do not have any ideal where to start. I am stopping with inetd.conf issue . I hope any one can help me or show me good document. Thanks edit
193 new active 2003 Mar anonymous 2016 Nov anonymous 3 2 Seting MIME types Our cvs repository contains various binary files, inclduing some {quote: PowerPoint} files. When browsing the repository cvstrac uses the MIME type text/plain for .ppt files. It would be nice to be able to configure this some how. edit
151 code active 2002 Dec anonymous 2017 Mar drh 4 1 html-encodes swedish chars in wiki pages with mozilla I tried to edit wikipages containing swedish characters with latest mozilla and when i previewed and submitted the changes, all these characters where encoded with amp-hash-asciicode... Editing with IE was ok, tho. Is it mozilla, me or cvstrac?. edit
813 code active 2009 Nov anonymous 2017 Mar cpb 4 4 Sorting reports with date columns The reports do all have column headers that work as hyperlinks and trigger a re-ordering of the report's records. That's a nice feature if used with the severity, priority and status columns, because usually the color of the records is only based on one of these columns. It doesn't work that smoothly with the date columns _origtime_ and _changetime_. Especially if you use sdate() to format the date columns. The way how it is implemented (and the simplicity of this has my admiration) applies the sorting to the formatted columns. rptview?rn=1&order_by=5&order_dir=ASC If you have more recent dates including changes of today you might have after the sort by the sdate() column: | 13:07 | A Ticket of today | | 2007 Nov | Ticket from 2 years back | | 2007 Oct | another Ticket | | Nov 05 | From this month | | Jun 18 | 4 months back | Check the report 9 to see this at work. At the moment I would need to guess if a report might use with "sorted by date columns", avoid sdate() for this report and use instead a "sortable" date like ISO date. But "sortable" has always the drawback to be longer, because the year must be always included. Would it be possible to use a derived sdate() function that includes a hidden sort part, something like: <span class="invisible">2009-11-12 13:07:23</span> 13:07 But that wouldn't be nice for older browsers that are not able to hide content. The other approach would be to change the generic way how the hyperlinks for the column headers are build. In that case the procedure that generates the header row needs to have some knowledge on the type of the selected columns. Seems to be complicated. edit
461 code active 2005 Aug cpb 2017 Mar cpb 4 4 can't upload attachments with weird filenames Uploading a filename like "test%%2526^*(++%2523.pl" as a Wiki attachment generates the following error message: <html> <p>The following error has occurred:</p> <blockquote>Attachment information is missing from the query content</blockquote> <p>Query parameters:<p> all_login = Ah1Pk_EMoQiUwMxAJiC_UG <br /> d = <br /> f = <br /> f:bytes = 0 <br /> f:filename = test%&^*(++%23.pl <br /> f:mimetype = application/x-perl <br /> ok = Add Attachment <br /> tn = WikiIndex <br /> </html> edit
719 code active 2007 Sep anonymous 2017 Mar cpb 4 3 Quote cleanup still not as advertised "Text is stripped of all single-quotes and backslashes before it is substituted, so if the substitution is itself enclosed in single-quotes, it will always be treated as a single token by the shell." Having a few blank or odd spots in the text is an OK compromise. That sure doesn't sound like all the text is going to be gone though. An ordinary user of cvstrac will likely use single quotes or back slash occasionally and then wonder why the text is not in the email message. OK I can get the description and remarks text out of the DB but only the wiki version. Not so nice as HTML! SO: 1) The warning is wrong. Maybe it should also be noted in the hints what want show up via notify. 2) Let(no single quote:)s add forward quote as another issue too. Ya know the one right below the tilde~ on most keyboards. If that is in the description or remarks then my email script doesn't even get called. edit
674 code active 2006 Dec anonymous 2017 Mar cpb 4 4 64 bit issues Hi guys, Thank you for this great software. I've installed it on a Debian/Etch system with sucess. It's working nearly fine in CGI mode with suExec module enabled. It seems that it is not possible to edit the tickets. I've tried with different users with full rights. When trying to edit the tickets, I get a server error with the following error logs in /var/log/apache2/error.log : index.cgi: line 2: 18143 Segmentation fault /usr/bin/cvstrac cgi /var/www/cvstrac server{linebreak} [Tue Dec 05 17:02:14 2006] [error] [client 130.104.38.125] Premature end of script headers: index.cgi, referer: https://www.pameseb.be/cvstrac/index.cgi/rptview?rn=1{linebreak} I've tried to use the cvstrack server directly to see if it was CGI related. The result is not better. When in server mode and trying to edit a ticket, I receive an empty page ... Any help would be greatly appreciated because at the moment, the system is not functionnal. I can provide much more details if needed. I can provide a web access to the installation if needed. Thanks in advance, Didrik edit
812 code new 2009 Oct anonymous 2017 Mar cpb 1 3 Tarball of cvstrac 2.0.1 STILL crashes on Snow Leopard (OS X 10.6) From #808 it was determined that _sqlite3_free_ needed called instead of _free_ in some cases for file db.c #808 notes that "This was fixed previously in [969] (almost two years ago)." However, the tarball download of version 2.0.1 of cvstrac does NOT seem to contain the fixes from [969], therefore a compiled version of this code of cvstrac on MacOS 10.6.1 will still crash with malloc errors. Can the tarball be updated to include the latest fixes? edit
846 event active 2020 Jun anonymous 2020 Jun anonymous 4 3 dupreez.peter Need assistance. Only a test edit
847 code active 2020 Jun anonymous 2020 Jun anonymous 4 3 96 INLAY Make me blades edit
845 doc active 2020 Jun anonymous 2020 Oct anonymous 4 3 dupreez.peter Need assistance with edit
849 code active 2022 Aug anonymous 2022 Aug cpb 4 3 test test edit
850 code active 2023 Mar anonymous 2023 Mar cpb 4 3 This is test bug abc edit
107 new active 2002 Aug rse Jan 26 rse 5 4 Timeline should trim commit log message to a reasonable size My check-in [178] shows up as a full message on the timeline. This destroys the compact and brief layout of the timeline page IMHO. Because it is fully valid to use full-page commit log messages on check-in, I think CVStrac should trim commit log messages to a reasonbale size for the timeline page. "Reasonable" perhaps means to first cut off everything after the first "\n\n" (empty line) to get just the first paragraph and if this paragraph is still longer than approx. 200 bytes I would trim it to the first 200 bytes. Finally after the trimming a " [...]" should be appended to clearly show that the message was trimmed. edit
Rows: 166