- 06 Apr, 2021 1 commit
-
-
Simon authored
-
- 02 Apr, 2021 1 commit
-
-
Guillaume Quintard authored
-
- 01 Apr, 2021 1 commit
-
-
Guillaume Quintard authored
-
- 31 Mar, 2021 2 commits
-
-
Dridi Boukelmoune authored
-
Dridi Boukelmoune authored
Refs #3565
-
- 30 Mar, 2021 4 commits
-
-
Poul-Henning Kamp authored
-
Dridi Boukelmoune authored
-
Dridi Boukelmoune authored
-
Dridi Boukelmoune authored
-
- 29 Mar, 2021 13 commits
-
-
Poul-Henning Kamp authored
See Also: #3557
-
Poul-Henning Kamp authored
Also compare in the exact resp.body, because now we can.
-
Poul-Henning Kamp authored
-
Nils Goroll authored
-
Nils Goroll authored
This is a TTFB optimization: I noticed that, when streaming, we would only sometimes send headers while waiting for more body data from the backend, depending on the amount of body data read ahead when reading the backend headers. The flush_head debug bit would avoid the delay, but it leads to an extra flush in all cases, even when we could send some body data down the line with the headers. This trivial patch peeks into the busy object to determine if the first ObjWaitExtend() is likely to block and, if so, issues a flush, which will cause the headers to be sent in addition to any initial data VDPs might have to send.
-
Dridi Boukelmoune authored
The assertion that the stale objcore of a conditional fetch cannot be failed unless it was streaming is incorrect. Between the moment when we grab the stale objcore in HSH_Lookup and the moment we try to use it after vcl_backend_response, the backend fetch may have completed or failed. Instead, we need to treat an ongoing fetch and a failed fetch as separate checks since the latter may happen with or without a boc.
-
Poul-Henning Kamp authored
-
Poul-Henning Kamp authored
-
Poul-Henning Kamp authored
-
Poul-Henning Kamp authored
-
Poul-Henning Kamp authored
-
Poul-Henning Kamp authored
-
Poul-Henning Kamp authored
-
- 27 Mar, 2021 8 commits
-
-
Nils Goroll authored
Please relace the vtree.h #pragma change of this commit with a better solution once you have one. 777e03ba had the disadvantage that it requires all vmods to adjust their build, besides affecting more code than necessary. Under the assumption that the issue with VRBT is not acutally a violation of strict aliasing, but rather only a violation of gcc's pedantic "compatible lvalue rule", we silence gcc warnings rather than turning off strict aliasing. Note that this is more convenient, but also more dangerous, because we fail to catch any actual strict aliasing errors where vtree.h is used. This reverts commit 777e03ba.
-
Nils Goroll authored
Move the address strdup() to where the acl_e is allocated. This plugs an utterly insignificant leak for error returns from vcc_acl_entry().
-
Nils Goroll authored
Apologies for the slight back and forth, but I now think this is a better way to make clear to flexelint that we either keep an allocated acl_e or not. This partly undoes changes from bf854df6 To review, I recommend diff'ing to a3664ead Ref aff516b7 f65f7148
-
Nils Goroll authored
Ref aff516b7 f65f7148
-
Nils Goroll authored
-
Nils Goroll authored
-
Nils Goroll authored
We only ever insert copies into the tree in vcc_acl_add_entry()
-
Nils Goroll authored
note: struct acl_e could be compacted, but we prefer clarify over efficiency in VCC.
-
- 26 Mar, 2021 1 commit
-
-
Nils Goroll authored
-
- 25 Mar, 2021 7 commits
-
-
Nils Goroll authored
Reduces conflicts with a held-back PR
-
Nils Goroll authored
As of now, this makes no difference to the vcc-generated code.
-
Nils Goroll authored
-
Nils Goroll authored
Each their own variant, *sigh*
-
Nils Goroll authored
No idea why I just overlooked these
-
Nils Goroll authored
-
Nils Goroll authored
to match the character at position 0x2F '/'
-
- 24 Mar, 2021 2 commits
-
-
Dridi Boukelmoune authored
There's no point waiting for the backend to complain if we weren't able to properly send the backend request. Fixes #3556
-
Dridi Boukelmoune authored
The reason we expect here can be summarized as: this was a pipe transaction or an error occurred. This could be much simpler if we replaced enum sess_close with a struct stream_close instead. Refs dc5bddbd
-