Commit Briefs

dad63a1458 Stefan Sperling

CHANGES for 0.117 (tags/0.117)


8b5c397a98 Stefan Sperling

make 'got clone' set a got.conf default branch for fetching only, not sending

This will make 'got send' default to the work tree's current branch or the HEAD reference if not invoked in a work tree, rather than defaulting to sending the default branch found while cloning. Issue pointed out by op@ ok op@


1a201d3a3d Stefan Sperling

add a gotsysd regress target which rebuilds binaries in the VM

Makes it easier to keep the same VM running for manual testing while tweaking the code.


c7694689e6 Stefan Sperling

gotsysd regress: run make clean in the template directory, else build can fail

Without make clean we would end up running a template binary built on the host system, which won't run in case libc.so is out sync between host and VM.


bcfb12b454 Thomas Adam

regress: git: don't load global/home configs

When running git(1) commands as part of the test suite, disallow git(1) from loading either the global or home gitconfig file. This has the ability to mess with the test and how they're setup. Each test already sets its own config so this is the preferred option anyway. OK stsp@


a6c26ae828 Stefan Sperling

add a todo item for a bug observed in gotd


eb01f20d38 Stefan Sperling

better fix for long loop bug during commit coloring with offloading

Repaint already queued commits in the main process instead of removing them from the queue. Otherwise we may never converge to a state where all commits are marked as skipped. Instead such commits can inherit color keep and we'll keep on asking got-read-pack to paint more commits for a very long time. ok op@



8cc3cb2155 Stefan Sperling

do not let got-read-pack paint parent commits which are not in its pack file

If any parent commit of the current commit is missing, stop iterating and return control to the main process. Fixes a case where offloading to got-read-pack resulted in an effectively endless loop. ok op@


c46b1605c7 Stefan Sperling

improve selection of pack files for pinning in the open pack file cache

If two given pack files contain the same amount of requested commits, prefer the pack file which contains the larger amount of total objects to increase our chances of full history traversal. Also, require the first element of the list of requested commits to be present in the pack file since the caller will very likely try to process this commit immediately when we return. This fixes cases where offloading of commit coloring to got-read-pack was ineffective, causing us to take the slow path instead. ok op@


65f9252095 Stefan Sperling

hack for got_privsep_recv_painted_commits() to sort queue by color

ok op@


3cdc21f2c7 Stefan Sperling

sort commits not processed by got-read-pack properly

ok op@


fcf8f0b2dc Stefan Sperling

pass an exact copy of queued IDs to got-read-pack during commit coloring

ok op@


7997350f7c Stefan Sperling

add functions which transport object ID queues across the imsg layer

ok op@


75123e1584 Omar Polo

dedup copyrights

while here move millert@ to the authors too.


0ec5a98736 Omar Polo

gotwebd: avoid needless close

send_imsg() does a imsg_flush() too which will ensure the fd is sent over to the other process and close it on this side. If it fails we terminate anyway, so no need to close (doubly!) the fds.


9f39b41c67 Stefan Sperling

ensure that drops and skips get processed before keeps during commit coloring

This makes repainting parent commits unnecessary and seems to avoid cases where our coloring algorithm ran into worst case performance behaviour. ok op@


eed86c18e9 Stefan Sperling

remove code which repaints parent commits from commit coloring implementation

Two pack tests start failing. This will be fixed in the next commit. ok op@


2b8971e958 Omar Polo

gotwebd: avoid the pipe; forward the socket directly

Instead of keeping a pipe between the sockets and the gotwebd process, just forward the socket directly. This needs a little bit of care because once we enter process_request() the socket (possibly) no longer there, so on the socket side use the client_status as a flag to signal the caller that we're done. I hope to improve this in follow-ups. The motivation behind this is to avoid a (possible, but quite easily triggerable) race in the current code where the other side might process GOTWEBD_IMSG_REQ_DONE before draining the pipe, truncating the output page. ok stsp@


274e82cea6 Omar Polo

improve binary blob detection

just like the change just done in diff.git, consider a blob binary if it has any control character (except for the tabs) joint work with jtt@, ok stsp@


ada5f16249 Omar Polo

sync files from diff.git 543aed44ef221a67c386d967ec45907fb129194b

improve binary files detection instead of 'just' looking for NUL, consider a file binary if it has any control charcters (except our beloved tabs characters). joint work with jtt@, ok stsp@


470daf1e6c Johannes Thyssen Tishman

preserve bad symlinks across merges

ok stsp@ and op@


772d6dc483 Omar Polo

fix math in fcgi_send_response

We can't hit actually hit this because our buffers are way smaller than FCGI_CONTENT_SIZE (65536) but if we were, we'd end up sending wrong fastcgi packets. found with / ok stsp@


6ce74ad156 Stefan Sperling

the gotsys auth/hmac index must be globally unique, not unique per repository

tested by Mischa Peters


9a88519990 Stefan Sperling

fix gotsysd using wrong auth and hmac labels in the generated gotd.conf

Problem reported by Mischa Peters