Re: [fossil-users] cloning / opening fails on WinXP SP3
On Tue, Jun 18, 2013 at 3:22 AM, Stephan Beal wrote: > i can commit to preparing one this week, and can test i32 and x64 ubuntu. > More concretely... i haven't been able to sleep tonight (thanks to a valgrind warning ;) and will be taking most of the day off today, so i will be able to prepare one this afternoon (that's in 10 or 12 hours). i've still go the links to the commit checklist somewhere, and will just start from there unless i hear otherwise. -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] cloning / opening fails on WinXP SP3
i can commit to preparing one this week, and can test i32 and x64 ubuntu. (sent from a mobile device - please excuse brevity, typos, and top-posting) - stephan beal http://wanderinghorse.net On Jun 18, 2013 2:15 AM, "Richard Hipp" wrote: > There is a fresh build of Fossil for windows at: > > http://www.fossil-scm.org/tmp/fossil-20130614-win.zip > > Please download that new version. Run "fossil rebuild" on the original > repo. Then try to clone using the new version as both client and server. > If you are still having problems after that, please send following email to > this mailing list. Tnx. > > (Aside to core committers: We really, really need to do a new release, > guys.) > > -- > D. Richard Hipp > d...@sqlite.org > ___ > fossil-users mailing list > fossil-users@lists.fossil-scm.org > http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users > > ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] Windows newb questions
Hi Richard, > Help me to understand: Are you starting a new project? If you are > starting to work on an existing project, where did you get the files to > work on if you didn't do the "source control prep" first? > In order to test Fossil, I essentially did the following: 1) created a folder for repositories (c:\fossil\repositories), and used fossil new to create a repo called test.fossil 2) went to the root of an existing Visual Studio project, and did a fossil open c:\fossil\repositories\test.fossil, saw the _FOSSIL_ file was created and 3) did fossil add . to add the existing files to the new repository At that point I saw all the debug detritus going into the repo, so I played with deleting and ignoring things. :) So anyway, that initial prep for an existing project has to be done with Git etc. as well. However at that point I wasn't sure if I have to prep each working session by opening the repo, working/committing, and then closing at the end. This guy http://ronperrella.blogspot.com/2012/12/fossil-first-steps.html says Fossil "doesn't like it" if the repository isn't open, and while he also mentions that there's little reason to close a repo, there's also little reason to think he must know what he's talking about since that information isn't on the official site, hence the question. > If you want to start the "source control prep" after the fact, that fine. > The only danger is file overwriting. Note also the --keep option to > "fossil open" which avoids all file overwriting. > That fully answers the question, thanks. > > But if you make changes to some version of code you got "out of band" and > then try to check it into fossil - are you sure you are checking it into > the right branch? Are you sure nobody else has changed it in the > meantime. There are lots of reasons not to do it that way. Just open the > checkout from fossil first, then start editing, and you will avoid lots of > potential problems and complications. > > >> Related to that, is it necessary to close the repository when you're >> done, or is it OK to leave it "open" between sessions, reboots, etc.? I >> just want to make sure I understand the workflow to recommend and why. >> > > Leave sessions open. There is hardly ever a good reason to close them. > Sitting here typing this, I cannot think of even one reason why you would > ever want to run "fossil close". (You will notice, btw, that "fossil > close" is not on the list of commonly used commands that appear when you > type "fossil help". ) > This is also good information, since it's not clear from a newbie perspective. > > A shared drive will work, in theory, assuming file locking works on the > shared drive. (Network filesystems are notoriously buggy in that > respect.) Performance won't be optimal, but will probably be good enough. > Hmm. I would hate to rely on Windows then if there's a reasonable possibility of corruption that way. > > > >> Would the preferred method be creating the Windows service to share a >> number of .fossil repositories, and then people can clone and autosync to >> the various URLs presented by the service? Thank you, >> >> > My opinion of the preferred solution is to run Fossil from CGI on a Linux > box. I'm guessing you aren't going to go for that solution, so my second > choice would be to run Fossil as a windows service someplace. I'd be happy to do a Linux box if I had the option, but I probably don't. So it sounds like in a pure-Microsoft environment, the windows service option is preferred. Thanks again, Pete ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] cloning / opening fails on WinXP SP3
There is a fresh build of Fossil for windows at: http://www.fossil-scm.org/tmp/fossil-20130614-win.zip Please download that new version. Run "fossil rebuild" on the original repo. Then try to clone using the new version as both client and server. If you are still having problems after that, please send following email to this mailing list. Tnx. (Aside to core committers: We really, really need to do a new release, guys.) -- D. Richard Hipp d...@sqlite.org ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] Windows newb questions
On Mon, Jun 17, 2013 at 7:47 PM, Pete Rihaczek wrote: > > Some questions: > > 1) Coming from using Mercurial and Git, I'm using to firing up Visual > Studio and getting to work without any source control prep. > Help me to understand: Are you starting a new project? If you are starting to work on an existing project, where did you get the files to work on if you didn't do the "source control prep" first? > If the Fossil repository isn't open at the root of my project (i.e. no > _FOSSIL_ file) and I then open it to commit changes I made while it was > closed, it will first prompt to overwrite the changed files. That makes > sense, and I can decline the overwrites and then commit, but it would seem > that the desired workflow would be to open the repository before starting > to work. I even found a blog entry that claims "Fossil doesn't like it" if > you don't do that. Is there any threat beyond perhaps accidentally > overwriting work? > If you want to start the "source control prep" after the fact, that fine. The only danger is file overwriting. Note also the --keep option to "fossil open" which avoids all file overwriting. But if you make changes to some version of code you got "out of band" and then try to check it into fossil - are you sure you are checking it into the right branch? Are you sure nobody else has changed it in the meantime. There are lots of reasons not to do it that way. Just open the checkout from fossil first, then start editing, and you will avoid lots of potential problems and complications. > Related to that, is it necessary to close the repository when you're done, > or is it OK to leave it "open" between sessions, reboots, etc.? I just > want to make sure I understand the workflow to recommend and why. > Leave sessions open. There is hardly ever a good reason to close them. Sitting here typing this, I cannot think of even one reason why you would ever want to run "fossil close". (You will notice, btw, that "fossil close" is not on the list of commonly used commands that appear when you type "fossil help". ) > > 2) I played with cloning a repository locally on the same drive, and > autosync worked as expected. But of course I don't want to host the master > repository for a project on my local drive, lest the drive fail. What would > be the preferred method for sharing multiple repositories using a shared > Windows fileserver? Simply throwing the files on a shared drive doesn't > seem right due to locking and contention issues(?). > A shared drive will work, in theory, assuming file locking works on the shared drive. (Network filesystems are notoriously buggy in that respect.) Performance won't be optimal, but will probably be good enough. > Would the preferred method be creating the Windows service to share a > number of .fossil repositories, and then people can clone and autosync to > the various URLs presented by the service? Thank you, > > My opinion of the preferred solution is to run Fossil from CGI on a Linux box. I'm guessing you aren't going to go for that solution, so my second choice would be to run Fossil as a windows service someplace. -- D. Richard Hipp d...@sqlite.org ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
[fossil-users] Windows newb questions
Greetings All, We're looking at using Fossil due to its support for ticketing (among other things), and I'm the designated point man to figure it all out and teach the rest of the lazy SOBs -- I mean developer team -- how it all works. :) Looks like a very nice bit of work, from what I've seen so far. ;) I confess that the "up and running in 5 minutes" tutorial took me considerably longer since I'm on Windows and had to work a few things out that weren't crystal clear. Perhaps incorporating some lessons learned into the "5 minute" intro might save other Windows users some time. First thing I wanted to was convert the .gitignore file I use with Visual Studio to Fossil's format. This is a typical one: https://github.com/github/gitignore/blob/master/VisualStudio.gitignore Since I don't think glob format includes a comment symbol, I just ran things together like so: *.suo *.user *.sln.docstates */[Bb]in/ */[Oo]bj/ *_i.c *_p.c *.ilk etc. That seems to work even without putting trailing asterisks on folders, e.g. */[Bb]in/*. Then you run into Windows quirks, such as the fact that you can't create a folder starting with a . from within the Explorer GUI, so creating a .fossil-settings folder at the project root and then a .ignore-glob file in that has to be done via the command line. So far so good. But then I decided I wanted to globally set crnl-glob to "*" to avoid having to type --no-warnings at every check-in, but following what I could find online only resulted in Fossil telling me I was being a knucklehead by repeatedly responding with the usage hint. It turns out that an asterisk in single quotes does the trick on Windows, i.e. fossil settings crnl-glob '*' --global. FWIW. Some questions: 1) Coming from using Mercurial and Git, I'm using to firing up Visual Studio and getting to work without any source control prep. If the Fossil repository isn't open at the root of my project (i.e. no _FOSSIL_ file) and I then open it to commit changes I made while it was closed, it will first prompt to overwrite the changed files. That makes sense, and I can decline the overwrites and then commit, but it would seem that the desired workflow would be to open the repository before starting to work. I even found a blog entry that claims "Fossil doesn't like it" if you don't do that. Is there any threat beyond perhaps accidentally overwriting work? Related to that, is it necessary to close the repository when you're done, or is it OK to leave it "open" between sessions, reboots, etc.? I just want to make sure I understand the workflow to recommend and why. 2) I played with cloning a repository locally on the same drive, and autosync worked as expected. But of course I don't want to host the master repository for a project on my local drive, lest the drive fail. What would be the preferred method for sharing multiple repositories using a shared Windows fileserver? Simply throwing the files on a shared drive doesn't seem right due to locking and contention issues(?). Would the preferred method be creating the Windows service to share a number of .fossil repositories, and then people can clone and autosync to the various URLs presented by the service? Thank you, Pete ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] Importing git copies and renames
On Mon, Jun 17, 2013 at 3:56 PM, Isaac Jurado wrote: > Hello, > > I finally came up with something to make import work with: > > git fast-export -M -C ... > > Below is the tentative patch that seems to work on my tests. The good > thing is that even though Fossil does not follows history across > renames, at least now the rename operation is visible and both source > and destination timelines are now linked. For repositories imported > from git, that is. > Thanks for the patch! These changes are in a branch ( www.fossil-scm.org/fossil/timeline?r=git-import-with-rename). I don't have good test cases for git imports, so I'll depend on others to let me know when this branch is ready to merge into trunk. -- D. Richard Hipp d...@sqlite.org ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
[fossil-users] Importing git copies and renames
Hello,
I finally came up with something to make import work with:
git fast-export -M -C ...
Below is the tentative patch that seems to work on my tests. The good
thing is that even though Fossil does not follows history across
renames, at least now the rename operation is visible and both source
and destination timelines are now linked. For repositories imported
from git, that is.
Please accept my apologies if I failed to follow the conventions.
Comments and testing will be very welcomed.
Index: src/import.c
==
--- src/import.c
+++ src/import.c
@@ -107,10 +107,11 @@
gg.nMerge = 0;
for(i=0; iisFrom==0 ) continue;
fossil_free(pFile->zName);
fossil_free(pFile->zPrior);
+pFile->zPrior = 0;
fossil_free(pFile->zUuid);
*pFile = gg.aFile[--gg.nFile];
i--;
}
}else
@@ -651,51 +658,36 @@
zTo = rest_of_line(&z);
i = 0;
mx = gg.nFile;
nFrom = strlen(zFrom);
while( (pFile = import_find_file(zFrom, &i, mx))!=0 ){
-if( pFile->isFrom==0 ) continue;
pNew = import_add_file();
pFile = &gg.aFile[i-1];
if( strlen(pFile->zName)>nFrom ){
pNew->zName = mprintf("%s%s", zTo, pFile->zName[nFrom]);
}else{
- pNew->zName = fossil_strdup(pFile->zName);
+ pNew->zName = fossil_strdup(zTo);
}
pNew->isExe = pFile->isExe;
pNew->isLink = pFile->isLink;
pNew->zUuid = fossil_strdup(pFile->zUuid);
pNew->isFrom = 0;
}
}else
if( memcmp(zLine, "R ", 2)==0 ){
- int nFrom;
import_prior_files();
z = &zLine[2];
zFrom = next_token(&z);
zTo = rest_of_line(&z);
i = 0;
- nFrom = strlen(zFrom);
while( (pFile = import_find_file(zFrom, &i, gg.nFile))!=0 ){
if( pFile->isFrom==0 ) continue;
-pNew = import_add_file();
pFile = &gg.aFile[i-1];
-if( strlen(pFile->zName)>nFrom ){
- pNew->zName = mprintf("%s%s", zTo, pFile->zName[nFrom]);
-}else{
- pNew->zName = fossil_strdup(pFile->zName);
-}
-pNew->zPrior = pFile->zName;
-pNew->isExe = pFile->isExe;
-pNew->isLink = pFile->isLink;
-pNew->zUuid = pFile->zUuid;
-pNew->isFrom = 0;
-gg.nFile--;
-*pFile = *pNew;
-memset(pNew, 0, sizeof(*pNew));
+pFile->zPrior = pFile->zName;
+pFile->zName = fossil_strdup(zTo);
+pFile->isFrom = 0;
}
- fossil_fatal("cannot handle R records, use --full-tree");
}else
if( memcmp(zLine, "deleteall", 9)==0 ){
gg.fromLoaded = 1;
}else
if( memcmp(zLine, "N ", 2)==0 ){
--
Isaac Jurado
"The noblest pleasure is the joy of understanding."
Leonardo da Vinci
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] cloning / opening fails on WinXP SP3
On Mon, Jun 17, 2013 at 04:40:09PM +0200, Michai Ramakers wrote: > On 17 June 2013 15:36, Richard Hipp wrote: > checksum mismatch on artifact 15: wanted > 91058d6d3dd1df16e04942a59bc970c7bcc04b61 but got > da39a3ee5e6b4b0d3255bfef95601890afd80709 That's an empty artifact. Joerg ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] Fossil, proxies and SSL, revisited
On Mon, Jun 17, 2013 at 10:59 AM, Lluís Batlle i Rossell wrote: > On Mon, Jun 17, 2013 at 10:56:03AM -0400, Richard Hipp wrote: > > On Mon, Jun 17, 2013 at 10:48 AM, Brandon Invergo >wrote: > > > > > > > > However, when I attempt to sync my Fossil repositories, I receive the > > > following error: > > > > > > $ fossil sync https://user:p...@repos.invergo.net/reponame > > > via proxy: http://proxy.company.com:8080 > > > fossil: server says: 503 Service Unavailable: 0 > > > > > > > I don't know that much about proxies, since I never personally need to > deal > > with them. But maybe they should be disabled for HTTPS? > > To use https over an http proxy, fossil should use a CONNECT http proxy > command. After success on that, it's like a normal direct connection to the > remote https server, and all TLS happens without the proxy understanding a > word. > I don't have access to a proxy for testing purposes. Can somebody (who has signed a CLA) suggest a patch to implement this? (CLA available here: http://www.fossil-scm.org/fossil/doc/trunk/www/copyright-release.pdf) -- D. Richard Hipp d...@sqlite.org ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] Fossil, proxies and SSL, revisited
On Mon, Jun 17, 2013 at 10:56:03AM -0400, Richard Hipp wrote: > On Mon, Jun 17, 2013 at 10:48 AM, Brandon Invergo wrote: > > > > > However, when I attempt to sync my Fossil repositories, I receive the > > following error: > > > > $ fossil sync https://user:p...@repos.invergo.net/reponame > > via proxy: http://proxy.company.com:8080 > > fossil: server says: 503 Service Unavailable: 0 > > > > I don't know that much about proxies, since I never personally need to deal > with them. But maybe they should be disabled for HTTPS? To use https over an http proxy, fossil should use a CONNECT http proxy command. After success on that, it's like a normal direct connection to the remote https server, and all TLS happens without the proxy understanding a word. ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] Fossil, proxies and SSL, revisited
On Mon, Jun 17, 2013 at 10:48 AM, Brandon Invergo wrote:
>
> However, when I attempt to sync my Fossil repositories, I receive the
> following error:
>
> $ fossil sync https://user:p...@repos.invergo.net/reponame
> via proxy: http://proxy.company.com:8080
> fossil: server says: 503 Service Unavailable: 0
>
I don't know that much about proxies, since I never personally need to deal
with them. But maybe they should be disabled for HTTPS? What happens if
you recompile with this patch:
Index: src/url.c
==
--- src/url.c
+++ src/url.c
@@ -326,11 +326,11 @@
if( zProxy==0 || zProxy[0]==0 || is_truth(zProxy) ){
zProxy = fossil_getenv("http_proxy");
}
}
if( zProxy && zProxy[0] && !is_false(zProxy)
- && !g.urlIsSsh && !g.urlIsFile ){
+ && !g.urlIsSsh && !g.urlIsFile && !g.urlIsHttps ){
char *zOriginalUrl = g.urlCanonical;
char *zOriginalHost = g.urlHostname;
char *zOriginalUser = g.urlUser;
char *zOriginalPasswd = g.urlPasswd;
unsigned uOriginalFlags = g.urlFlags;
--
D. Richard Hipp
d...@sqlite.org
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
[fossil-users] Fossil, proxies and SSL, revisited
[sorry if this is received twice; I originally sent it before I received my list membership confirmation, so I think the original was discarded] Hello, I'm having problems syncing with my Fossil repositories via HTTPS from behind a proxy. I have searched the list archives and I saw a similar topic from 2010 that did not show any possible resolution for me. As a caveat, I profess that I am generally ignorant of the inner workings of proxies. I'm hosting my repositories from a CGI script on a shared host. I have a self-signed SSL certificate for the subdomain that hosts the repos. From home, I can sync via HTTPS without problems. However, at work I am behind a particularly pesky proxy. Nevertheless, I can visit the repos' web pages via HTTPS in a browser despite the proxy. Similarly, I can, for example, from the command line use wget to fetch a file via HTTPS From the server through the proxy, and I can confirm that it is really going via HTTPS because it complains about my self-signed certificate. However, when I attempt to sync my Fossil repositories, I receive the following error: $ fossil sync https://user:p...@repos.invergo.net/reponame via proxy: http://proxy.company.com:8080 fossil: server says: 503 Service Unavailable: 0 If I set the proxy to https://proxy.company.com:8080 (probably stupid, I know, but let's try it anyway), I get the following: fossil: SSL: cannot connect to host proxy.company.com:8080 (unknown protocol) I have built Fossil myself, and I was sure to build in SSL support. Just now I logged into another machine that I have an account on that is not behind a firewall and I confirmed that I can clone a repository via HTTPS, so the 503 error does not indicate some temporary failure in service. So, I'm at a bit of a loss on what other steps to try to get around this. Does anyone have any suggestions? Thanks for your help! -- Brandon Invergo http://brandon.invergo.net pgpn2taJr8BsO.pgp Description: PGP signature ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] cloning / opening fails on WinXP SP3
On Mon, Jun 17, 2013 at 10:40 AM, Michai Ramakers wrote:
> On 17 June 2013 15:36, Richard Hipp wrote:
> >
> >> 3) notice the cloned repo is too small; roughly 19 MB instead of
> >> expected 311 MB as on the http host (this may be irrelevant - no
> >> idea.)
> >
> > Where are you seeing this 311MB number? See
> > http://www.fossil-scm.org/fossil/stat for the current repo size on the
> host.
> > Looks like 29MB. 19MB might be a reasonable clone.
>
> Thank you, I didn't know the 'stat' command. Result: 311085056 bytes
> (311.1MB)
>
I misread your original post. I thought you were trying to clone the
Fossil self-hosting repository.
>
> BTW, I didn't mean to imply that the 'repo is too small' really; I
> cannot judge that. What I meant was that it seemed too small on first
> sight, of course.
>
> > What does "fossil ui" do? Do you get the entire repository?
>
> On the http host, providing the repo file as argument failed ("Error
> showing url: Operation not supported" - I tried setting web-browser to
> full path of firefox, which failed as well). 'fossil server -P 8080
> /fossils/my_fossil.fossil' worked, and showed the full history and
> full fileset.
>
> On the http client, providing the clined 19MB repo file as argument: I
> see the timeline with only the initial checkin as in the original
> mail. I see no files (which would correspond to the set of files in
> the repo at the time of that checkin).
>
> > What does "fossil test-integrity" say?
>
> On the http host, locally:
>
> $ f test-integrity /fossils/my_fossil.fossil
> 2066 non-phantom blobs (out of 2066 total) checked: 0 errors
> $
>
> On the http client, after clone:
>
> C:\proj\repo>f test-integrity my_fossil.fossil
> checksum mismatch on artifact 15: wanted
> 91058d6d3dd1df16e04942a59bc970c7bcc04b61 but got
> da39a3ee5e6b4b0d3255bfef95601890afd80709
> checksum mismatch on artifact 16: wanted
>
This very clearly indicates a problem with the clone. I don't know what
might be going wrong. If you are able to send me a copy of the repository
you are failing to clone (or just URL so that I can clone it myself) I can
try to reproduce the problem.
This is not a problem that I recall ever seeing before.
> 216c9bd264a36f9f4c9a88e2c42b0d30f3e48026 but got
> da39a3ee5e6b4b0d3255bfef95601890afd80709
> skip phantom 18 66e7811833dc5230df72a98c5edf840b9001b084 checksum
> mismatch on artifact 547: wanted
> 37a1de3e9441fcaf8371acb4f5ad216bf1ff9712 but got
> 2cfab7c90f8ad60277cf63c3c4bcfbcdb2ea85d7
> 546 non-phantom blobs (out of 547 total) checked: 3 errors
> C:\proj\repo>
>
> Michai
> ___
> fossil-users mailing list
> fossil-users@lists.fossil-scm.org
> http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
>
--
D. Richard Hipp
d...@sqlite.org
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] cloning / opening fails on WinXP SP3
On 17 June 2013 16:41, Eduardo Morras wrote: > > Did you rebuild after clone? Try fossil rebuild my_fossil.fossil --analyze > > Note that clone doesn't copy user information. I think, correct me if I'm > wrong, it doesn't copy information that can be rebuild from artifacts, like > indexes. Tried, same result (only _FOSSIL_ file after 'fossil open') Michai ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] cloning / opening fails on WinXP SP3
On 17 June 2013 15:36, Richard Hipp wrote: > > Where are you seeing this 311MB number? See > http://www.fossil-scm.org/fossil/stat for the current repo size on the host. > Looks like 29MB. 19MB might be a reasonable clone. Just to clarify: I am not trying to clone the Fossil source repository, but my own repo, I hope that was clear. (However, as a test, I also cloned the Fossil source repository onto the Win32 host, which went fine.) Michai ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] cloning / opening fails on WinXP SP3
On Mon, 17 Jun 2013 15:20:34 +0200
Michai Ramakers wrote:
> Hello,
>
> this is the first time I try Fossil on a Win32 system, so I'm guessing
> I do something wrong. Here is what I did, and what fails:
>
> 1) download fossil executable from fossil-scm.org; this is 1.25 [d2e07756d9]
> 2) in dir '\some_where\repo', clone
> http://michai@otherhost_on_lan/my_fossil my_fossil.fossil
>
> (otherhost_on_LAN is running fossil 1.25 [c7133bd79d], amd64 netbsd)
>
> 3) notice the cloned repo is too small; roughly 19 MB instead of
> expected 311 MB as on the http host (this may be irrelevant - no
> idea.)
> 4) in dir '\some_where\src', try to open the cloned repo anyway
> 5) notice only a _FOSSIL_ file is created, nothing else
> 6) 'fossil stat' in 'src' shows '[f34e9990ce] initial empty check-in
> (user: michai, tags: trunk)' (which corresponds to the oldest timeline
> entry of that repo)
>
> I tried cloning another repo from the same host, which went fine.
>
> Fossil as http server is running from inetd on 'other_host_on_LAN'.
>
> The Win32 host is running Windows XP pro SP3.
>
> Running 'fossil clone' locally on 'other_host_on_LAN' works fine for
> the same repo (with cloned repo being around 466 MB, fwiw).
>
> Running 'fossil clone' on another amd64 netbsd, cloning the same repo
> on 'other_host_on_LAN' works too; the fossil version on that host
> happened to be 1.24 [8d758d3715] .
>
> Perhaps worth knowing, or not: while working on 'other_host_on_LAN', I
> had trouble adding or checking in (cannot remember) files with strange
> characters in them - somehow managed anyway, but later removed these
> files again ('fossil addremove'). In between the initial checkin I saw
> cloned on the Win32 host and the moment of adding these files, a few
> other checking had happened (which perhaps means these are 2 unrelated
> things - no idea).
>
> So... that's all; any ideas are welcome. I don't currently have a
> mingw32/cygwin/whatever native build environment on the Win32 host; if
> it's relevant, I could install it and try new fossil-versions from
> source.
Did you rebuild after clone? Try fossil rebuild my_fossil.fossil --analyze
Note that clone doesn't copy user information. I think, correct me if I'm
wrong, it doesn't copy information that can be rebuild from artifacts, like
indexes.
> Michai
--- ---
Eduardo Morras
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] cloning / opening fails on WinXP SP3
On 17 June 2013 15:36, Richard Hipp wrote:
>
>> 3) notice the cloned repo is too small; roughly 19 MB instead of
>> expected 311 MB as on the http host (this may be irrelevant - no
>> idea.)
>
> Where are you seeing this 311MB number? See
> http://www.fossil-scm.org/fossil/stat for the current repo size on the host.
> Looks like 29MB. 19MB might be a reasonable clone.
Thank you, I didn't know the 'stat' command. Result: 311085056 bytes (311.1MB)
BTW, I didn't mean to imply that the 'repo is too small' really; I
cannot judge that. What I meant was that it seemed too small on first
sight, of course.
> What does "fossil ui" do? Do you get the entire repository?
On the http host, providing the repo file as argument failed ("Error
showing url: Operation not supported" - I tried setting web-browser to
full path of firefox, which failed as well). 'fossil server -P 8080
/fossils/my_fossil.fossil' worked, and showed the full history and
full fileset.
On the http client, providing the clined 19MB repo file as argument: I
see the timeline with only the initial checkin as in the original
mail. I see no files (which would correspond to the set of files in
the repo at the time of that checkin).
> What does "fossil test-integrity" say?
On the http host, locally:
$ f test-integrity /fossils/my_fossil.fossil
2066 non-phantom blobs (out of 2066 total) checked: 0 errors
$
On the http client, after clone:
C:\proj\repo>f test-integrity my_fossil.fossil
checksum mismatch on artifact 15: wanted
91058d6d3dd1df16e04942a59bc970c7bcc04b61 but got
da39a3ee5e6b4b0d3255bfef95601890afd80709
checksum mismatch on artifact 16: wanted
216c9bd264a36f9f4c9a88e2c42b0d30f3e48026 but got
da39a3ee5e6b4b0d3255bfef95601890afd80709
skip phantom 18 66e7811833dc5230df72a98c5edf840b9001b084 checksum
mismatch on artifact 547: wanted
37a1de3e9441fcaf8371acb4f5ad216bf1ff9712 but got
2cfab7c90f8ad60277cf63c3c4bcfbcdb2ea85d7
546 non-phantom blobs (out of 547 total) checked: 3 errors
C:\proj\repo>
Michai
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] cloning / opening fails on WinXP SP3
On Mon, Jun 17, 2013 at 9:20 AM, Michai Ramakers wrote:
> Hello,
>
> this is the first time I try Fossil on a Win32 system, so I'm guessing
> I do something wrong. Here is what I did, and what fails:
>
> 1) download fossil executable from fossil-scm.org; this is 1.25
> [d2e07756d9]
> 2) in dir '\some_where\repo', clone
> http://michai@otherhost_on_lan/my_fossil my_fossil.fossil
>
> (otherhost_on_LAN is running fossil 1.25 [c7133bd79d], amd64 netbsd)
>
> 3) notice the cloned repo is too small; roughly 19 MB instead of
> expected 311 MB as on the http host (this may be irrelevant - no
> idea.)
>
Where are you seeing this 311MB number? See
http://www.fossil-scm.org/fossil/stat for the current repo size on the
host. Looks like 29MB. 19MB might be a reasonable clone.
> 4) in dir '\some_where\src', try to open the cloned repo anyway
> 5) notice only a _FOSSIL_ file is created, nothing else
> 6) 'fossil stat' in 'src' shows '[f34e9990ce] initial empty check-in
> (user: michai, tags: trunk)' (which corresponds to the oldest timeline
> entry of that repo)
>
What does "fossil ui" do? Do you get the entire repository?
What does "fossil test-integrity" say?
>
> I tried cloning another repo from the same host, which went fine.
>
> Fossil as http server is running from inetd on 'other_host_on_LAN'.
>
> The Win32 host is running Windows XP pro SP3.
>
> Running 'fossil clone' locally on 'other_host_on_LAN' works fine for
> the same repo (with cloned repo being around 466 MB, fwiw).
>
> Running 'fossil clone' on another amd64 netbsd, cloning the same repo
> on 'other_host_on_LAN' works too; the fossil version on that host
> happened to be 1.24 [8d758d3715] .
>
> Perhaps worth knowing, or not: while working on 'other_host_on_LAN', I
> had trouble adding or checking in (cannot remember) files with strange
> characters in them - somehow managed anyway, but later removed these
> files again ('fossil addremove'). In between the initial checkin I saw
> cloned on the Win32 host and the moment of adding these files, a few
> other checking had happened (which perhaps means these are 2 unrelated
> things - no idea).
>
> So... that's all; any ideas are welcome. I don't currently have a
> mingw32/cygwin/whatever native build environment on the Win32 host; if
> it's relevant, I could install it and try new fossil-versions from
> source.
>
> Michai
> ___
> fossil-users mailing list
> fossil-users@lists.fossil-scm.org
> http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
>
--
D. Richard Hipp
d...@sqlite.org
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
[fossil-users] cloning / opening fails on WinXP SP3
Hello,
this is the first time I try Fossil on a Win32 system, so I'm guessing
I do something wrong. Here is what I did, and what fails:
1) download fossil executable from fossil-scm.org; this is 1.25 [d2e07756d9]
2) in dir '\some_where\repo', clone
http://michai@otherhost_on_lan/my_fossil my_fossil.fossil
(otherhost_on_LAN is running fossil 1.25 [c7133bd79d], amd64 netbsd)
3) notice the cloned repo is too small; roughly 19 MB instead of
expected 311 MB as on the http host (this may be irrelevant - no
idea.)
4) in dir '\some_where\src', try to open the cloned repo anyway
5) notice only a _FOSSIL_ file is created, nothing else
6) 'fossil stat' in 'src' shows '[f34e9990ce] initial empty check-in
(user: michai, tags: trunk)' (which corresponds to the oldest timeline
entry of that repo)
I tried cloning another repo from the same host, which went fine.
Fossil as http server is running from inetd on 'other_host_on_LAN'.
The Win32 host is running Windows XP pro SP3.
Running 'fossil clone' locally on 'other_host_on_LAN' works fine for
the same repo (with cloned repo being around 466 MB, fwiw).
Running 'fossil clone' on another amd64 netbsd, cloning the same repo
on 'other_host_on_LAN' works too; the fossil version on that host
happened to be 1.24 [8d758d3715] .
Perhaps worth knowing, or not: while working on 'other_host_on_LAN', I
had trouble adding or checking in (cannot remember) files with strange
characters in them - somehow managed anyway, but later removed these
files again ('fossil addremove'). In between the initial checkin I saw
cloned on the Win32 host and the moment of adding these files, a few
other checking had happened (which perhaps means these are 2 unrelated
things - no idea).
So... that's all; any ideas are welcome. I don't currently have a
mingw32/cygwin/whatever native build environment on the Win32 host; if
it's relevant, I could install it and try new fossil-versions from
source.
Michai
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
