I've been fooling around with producing a longer piece of documentation using material from the tutorials, etc., using the docbook backend and "book" doctype. It's not working quite the way I'd expect. For example, it doesn't format the literal blocks correctly, and man page links aren't producing useful clickable links. Should I be going about this some other way? All I really want is automatically generated tables of contents and cross-references. --b. -

I have added this to GitWiki: check out http://git.or.cz/gitwiki/InterfacesFrontendsAndTools#cgit Hmmm... git has now 4 web interfaces (5 if counting gittracker, which works but gittracker repository shows empty): gitweb in Perl, git-php in PHP, wit in Python (is it actively developed? where it can be downloaded from?), gitarella in Ruby, now CGit in C... By the way, it is cgit, CGit or CGIt ;-p? -- Jakub Narebski Warsaw, Poland ShadeHawk on #git -

Your pseudo-algorithm is dubious: name = generate_cache_name(request); top: if (!exists(name)) { if (lock_cache(name)) { generate_cache(request, name); unlock_cache(name); } else { sched_yield(); goto top; } } else if (expired(name)) { if (lock_cache(name)) { generate_cache(request, name); unlock_cache(name); } } print_file(name); You really should have: if (!exists) { if (!lock) delay-and-repeat; /* RETEST ...

What happens if you have "[PATCH" without closing "]"? Does it work (and doesn't crash)? Does it strip anything? -- Jakub Narebski Warsaw, Poland ShadeHawk on #git -

The idea is to not store /etc in git directly, but use import/export scripts, which for example saves permissions and ownership in some file also tracked by git on import, and restores correct permissions on export. That is what I remember from this discussion. This of course means that you would have to write your own porcelain... What about mentioned in other email IsiSetup? -- Jakub Narebski Warsaw, Poland ShadeHawk on #git -

It should be "git clone --bare git://git.kernel.org/pub/scm/git/git.git" or "git clone --bare git://git.kernel.org/pub/scm/git/git.git git.git" True, the error messages of git-clone are bit cryptic and doesn't give us much information. Does there exist such repository? Perhaps WORKSFORME $ git --version git version 1.4.4.1 $ git clone --bare git://git.kernel.org/pub/scm/git/git.git remote: Generating pack... remote: Done counting 33587 objects. remote: Deltifying 33587 ...

Hi, This is what my libcurl libs looks like: ~% pkg-config --libs libcurl -lcurl -lcares -lidn -lssl -lcrypto -ldl -lz But the git Makefile assumes '-lcurl' will do. Bug identical to: https://trac.xiph.org/ticket/1092 # Han -- _/| VK |\_ // o\ Decision maker, n.: The person in your office who /o \\ || ._) was unable to form a task force before the music (_. || //__\ ...

I use stgit 0.11 1056:[gitweb/web!git]$ stg help usage: stg help <command> Ah. I'm sorry, I haven't noticed this. It is Hmmm... in stgit 0.11 "stg refresh --sign" once gave me Signed-off-by: The git commands StGit uses to perform operations automatically record changes in branches in reflog. What StGit does not provide is the "reason". You do use git-update-ref? -- Jakub Narebski Poland -

Yes, only for updating HEAD. The refs in refs/patches/<branch>/ are written directly. I initialy wanted to add patch history support using reflogs and added "git-update-ref -m ..." for the patch commits but I found slow the pushing operation a bit. Do you only want to track the reflogs for HEAD? -- Catalin -

Ooops, I haven't been clear enough. I meant that afer "stg commit; stg clean" I won't have any StGIT metadata, but I'd have git metadata in reflog. I'd like to have info in reflog for each command which changes head; for example "push", "refresh", perhaps "pop", "float", "uncommit". Just so I don't have long sequence of ref changes in reflog without description of said changes after some work with StGIT on branch. BTW. currently I use StGIT to manage a series of commits on ...

The real problem I have with that is you literally have to duplicate all sorts of functionality. I want to run "foo-status" in /etc and get something useful, but if /etc is not a git directory in and of itself then you have to duplicate most of "git-status" anyways. And the same applies to all the other commands. From what I can see of IsiSetup the tools for checking out, merging, modifying, cloning, etc are all much more limited and immature than the ones available through ...

>>> I actively use git to version, store and distribute an exim mail

Hi, Sorry about the missing subject. Make sure $EDITOR is executed as a command. Otherwise, _editor may fail if $EDITOR is set to something like 'editor', which is also a function within cg-Xlib. Signed-off-by: Clemens Buchacher <drizzd@aon.at> --- commit 942341e051fdcbb77a6abfbc58cf08ef8cab388d tree 96e8fd08cdc7c59b0d3d55d7cb5b4302db443aff parent b6a6e87cb3e1368ad0f78c18fdb6c29dde4ae83e author Clemens Buchacher <drizzd@aon.at> Sun, 10 Dec 2006 14:37:37 +0100 committer Clemens Buchacher ...

On Dec 10, 2006, at 10:06:14, Santi B

I've recently become somewhat interested in the idea of using GIT to store the contents of various folders in /etc. However after a bit of playing with this, I discovered that GIT doesn't actually preserve all permission bits since that would cause problems with the more traditional software development model. I'm curious if anyone has done this before; and if so, how they went about handling the permissions and ownership issues. I spent a little time looking over how GIT stores ...

Not historically at all. "git checkout" needed exact filenames in older versions. However, since about 1.4.4.1 or so, it now does the same filename expansion as "git add" etc does, which means that you can give it a directory name and it will check out everything under that directory, or you can give it a pattern, and it should glob it. But it sounds like you may have a slightly older version of git (the pathname matching really is fairly recent). Linus -

Right. Because every time you sleep, you sleep for 100 ms. That's why I was saying that you need to add polling to the thing. I don't know what the QT interfaces to asynchronous polling file descriptors are, but as long as you just blindly wait for 100ms whenever you run out of data, things will always suck. Using "fread()" hid this problem, because the thing would block in How about you just compare something simpler: git-rev-list | cat > /dev/null vs git-rev-list > ...

Why don't you use the pipe and standard read()? Even if you use "popen()" and get a "FILE *" back, you can still do int fd = fileno(file); and use the raw IO capabilities. The thing is, temporary files can actually be faster under Linux just because the Linux page-cache simply kicks ass. But it's not going to be _that_ big of a difference, and you need all that crazy "wait for rev-list to finish" and the "clean up temp-file on errors" etc crap, so there's no way it's a better ...

Why do you seem to be doing a "new" on every iteration inside the loop? Also, why do you have that strange FILE_BLOCK_SIZE thing, and in particular the "if (len < FILE_BLOCK_SIZE)" check? One thing that pipes vs files do is the blocking factor. Especially with older kernels, I _guarantee_ you that you'll only ever get 4kB at a time, so because of that "if (len < 64kB) break" thing, the only thing you're doing is to make sure things suck performance-wise, and you won't be reading ...

Of course is if (canceling) { deleteLater(); return; } Sorry, cut & paste typo ;-) -

Well, that's still just going slower than swapping. The reason loading things into memory is nice is that: - in the common case, you don't need to do anything else. - if the machine is low on memory, it can page things out just about as easily as you could write things to a file anyway. So don't worry too much about low-memory situations. Yes, there are cases where it's better to keep things in files and simply not have a big working set AT ALL, but if you keep something in a ...

That's because fread() will block until it gets all data. Did you actually ever try this with a uncached tree and did you compare what you got with a plain "read()". On older kernels, I guarantee that you get 4kB at a time for reads, even for a blocking pipe. Because we have bigger pipe buffers these days, it _may_ return 64kB at a time every time, but only if the writer is much faster than the reader. Based on the fact that you say that "read()" was ten times slower than ...

Yes it's right. Thanks! But it's still slow. Almost one second (985ms) to read the little qgit repo: $ ./qgit HEAD Found GNU source-highlight 2.5 count is <-1> count is <60169> count is <-1> count is <60505> count is <-1> count is <61462> count is <-1> count is <61911> count is <-1> count is <61392> count is <-1> count is <61880> count is <-1> count is <62009> count is <-1> count is <62549> count is <-1> count is <21354> count is <0> $ As a compare the temporary file ...

And as far as I can see it doe not use git credentials (user.name and And yet another one: better support for reflog, namely giving the "reason" i.e. the reflog message (like "stg push: <subject>", "stg refresh: <subject>", "stg pop: <subject>", "stg commit" etc.), like git-rebase, git-commit --amend and git-am (for example) does. P.S. The Vendor: field in stgit RPM has incorrectly Catalin Marinas <catalin.marinas@gmail.org> instead of Catalin Marinas ...

Hi, Okay, after thinking about it, I agree. merge-recursive is really the only user for such a diff. So, I do not think EMPTY_TREE would be useful. Ciao, Dscho -

The problem is that you forgot that the lines are indented when acting as a filter. How about this? It also contains the extra whitespace removal from the author name. diff --git a/builtin-shortlog.c b/builtin-shortlog.c index 7a2ddfe..3fc43dd 100644 --- a/builtin-shortlog.c +++ b/builtin-shortlog.c @@ -188,18 +188,25 @@ static void read_from_stdin(struct path_list *list) bob = buffer + strlen(buffer); else { offset = 8; - if (isspace(bob[-1])) + while (buffer + ...

Hi, From the current builtin-shortlog.c:128ff: if (!strncmp(oneline, "[PATCH", 6)) { char *eob = strchr(oneline, ']'); if (eob) { while (isspace(eob[1]) && eob[1] != '\n') eob++; if (eob - oneline < onelinelen) { onelinelen -= eob - oneline; oneline = eob; } } ...

s

Maybe I am missing something. It tells you about added or removed files (either binary or non-binary). I'd prepare a pair of practice patch files. "forward" has creation, "reverse" has addition. $ git diff-tree -p --binary fe142b3a >forward.patch $ git diff-tree -R -p --binary fe142b3a >reverse.patch $ git apply --summary --numstat forward.patch 50 20 git-cvsexportcommit.perl 145 0 t/t9200-git-cvsexportcommit.sh - - t/test9200a.png - - ...

Oh, I see now ("Fast access git-rev-list output..."). BTW, I just cannot reproduce that at all (on Linux): time { git rev-list --all > /tmp/ggg; cat /tmp/ggg >/dev/null; } tends to be somewhat slower than time git rev-list --all | cat >/dev/null Dunno. It's hard to assess on "small" repos, like kernel. They feel almost equally fast (maybe because qgit checks working directory too). popen(3) usually uses pipe(2). It's also awkward with regard to shell metacharacters and ...

Hi, This is exactly the kind of assumption which led to the bug to begin with. Ciao, Dscho -

Tested - result is the same, git-damon strace: [pid 6686] write(1, "00408419a453dc088b25b63ab1746d3d7e679caf686d refs/tags/v1.4.1.1\n", 64) = 64 [pid 6686] write(2, "fatal: ", 7) = 7 [pid 6686] write(2, "corrupted pack file ./objects/pack/pack-bc830a5d1c2efa3b727ef9af8faba13c3e719395.pack", 85) = 85 Options handling in git* is a nightmare. It accepts non-existant options. You -- Arkadiusz Mi

I do. I wanted to see if it reproduced but it didn't for me. -

The "but" in this sentence doesn't seem right to me, I would either use "although", or slightly better (IMHO) make it two sentences Tell the command to automatically stage files that have been modified and deleted. Note that files that you have not told git about are not included with this option. -- Alan Chandler http://www.chandlerfamily.org.uk -

regardless, I think "it" should become "them" ... it is tracking and commit them automatically. -- Alan Chandler http://www.chandlerfamily.org.uk -

OK, OK. Shortest I can manage then is: by using gitlink:git-add[1] to add new or modified files to the next commit. This adds a file's current contents only; run gitlink:git-add[1] again to include any subsequent changes. which may not be any improvement. --b. -

If gitweb produced cache-friendly headers, squid could definitely serve as an HTTP front-end ("HTTP accelerator" mode in squid talk). In fact, given kernel.org's slave1/slave2<->master setup, that's a pretty natural fit for caching files and/or cache-aware CGI output. You could even replace rsync to the slaves, if squid was serving as the front-end accelerator running on the slaves, communicating to the master. squid is smart enough to hold off a thundering herd, and only pulls single ...

It's impossible, Apache doesn't supply e-tag info to CGI programs. (it does supply HTTP_CACHE_CONTROL though apparently) You could probably do it via mod_perl. Jeff -

In the case of Perl scripts, it's not really the fork/exec overhead, but the Perl startup overhead that you want to try to optimize. But given your later statement (lots of spare cpu), this ends up just being a bit of a latency hit. In general, I think mod_perl has a much bigger impact when you have a database to connect to at startup. -

I/O - nonexistent; that stuff will be in memory. CPU - we have more CPU than you can shake a stick at, and it's 95+% idle. *NOT AN ISSUE*. -hpa -

Well, the idea (perhaps stupid idea: I don't know how caching engines / reverse proxy works) was that there would be caching engine / reverse proxy in the front (Squid for example) would cache results and serve it to rampaging hordes. But this caching engine has to ask gitweb if the cache is valid using "If-Modified-Since:" and "If-None-Match:" headers. What about the other idea, the one with raising expires to infinity for immutable pages like "commit" view for commit given by SHA-1? Even ...

It is up to the script (CGI or via mod_perl) to set the status to 304 and finish execution. Just setting the status to 304 does not forcefully end execution as you may want to cleanup, log, etc. cheers, martin -

You can use this attached shell script as a CGI script, to see precisely what information Apache gives you. You can even experiment with passing back headers other than Content-type (such as E-tag), to see what sort of results are produced. The script currently passes back both E-Tag and Last-Modified of a sample file; modify or delete those lines to suit Sorry, I was /assuming/ mod_perl would make this available. The HTTP header info is available to all Apache modules, but I confess ...

Sometimes it is easier to use ETags, sometimes it is easier to use Last-Modified:. Usually you can check ETag earlier (after calling git-rev-list) than Last-Modified (after parsing first commit). But some pages doesn't have natural ETag... Besides, because ETag is HTTP/1.1 we should provide and validate both. P.S. Any hints to how to do this with CGI Perl module? -- Jakub Narebski Poland -

Sure, if the proxies actually do the rigth thing (which they may or may I agree, but as mentioned, I think the _real_ problem tends to be the pages that don't act that way (ie summary pages, both at the individual project level and the top "all projects" level). Linus -

That is fine by me. We would benefit from an empty blob and an empty tree. -

merge-recursive and "diff against root" Linus -

I think this might be fragile in the presense of older implementation of git. A new git may operate fine without having "well known" objects in the repository, and would happily write result that points at them; an older git would find the resulting repository corrupt. We somehow need to force "well known" objects to be instantiated in the object database when something else points at them. I am not absolutely sure if keeping has_sha1_file() unaware of this magic is enough. -- >8 ...

I was wondering if we ever had any special case where we wanted the empty blob, and couldn't come up with any. Unlike the "tree diff" case, a "blob diff" will not actually do a "real diff" with a non-existing object, it special-cases it (and, for performance reasons, I think it really should). And all the blob-diff routines generally want to be able to take explicit data anyway, not just diffing two SHA1's (since we often don't have a SHA1 anyway - the working tree case). So while ...

Junio C Hamano wrote in I not necessarily agree; I think that symbolic links are sufficnetly rare that a bit more time spent to make the view better for end user (link only if target exists) is worth it. But... Here follows the implementation of this idea: first to read link target and show it in "tree" view _without_ hyperlink, then introduction of generic "object" view which does the verification and redirect to correct view accorting to the type of object, and last show link target ...

I am certainly in favor of the approach. Will look at the code later. Thanks. -

Junio C Hamano wrote in I not necessarily agree; I think that symbolic links are sufficnetly rare that a bit more time spent to make the view better for end user (link only if target exists) is worth it. But... Here follows the implementation of this idea: first to read link target and show it in "tree" view _without_ hyperlink, then introduction of generic "object" view which does the verification and redirect to correct view accorting to the type of object, and last show link target ...

Make symbolic link target in "tree" view into hyperlink to generic "object" view (as we don't know if the link target is file (blob) or directory (tree), and if it exist at all). Target of link is made into hyperlink when: * hash_base is provided (otherwise we cannot find hash of link target) * link is relative * in no place link goes out of root tree (top dir) Full path of symlink target from the root dir is provided in the title attribute of hyperlink. Currently symbolic link ...