| <<<Back 1 day (to 2017/06/10) | 20170611 |
sebras | tor8 (for the logs): so I have updated both -mini and -old to try to get them in synch | 04:44.13 |
| tor8: the important changes are: removing the app name/version number from the titlebar in -old (keeping in synch with -mini) | 04:44.46 |
| tor8: supporting running app on multiple devices using ANDROID_SERIAL | 04:45.12 |
| tor8: we did discuss the latter a while back and had different solutions, hopefully I merged them acceptably into one | 04:45.58 |
| tor8: finally I spent a lot of time trying to convince gradle to choose the APK versionCode based on ABI while keeping our existing numbering scheme | 04:47.03 |
| tor8: I think I pulled this off successfully, at least it appears to work. :) | 04:47.18 |
| tor8: so please have a look at the commits on -old and -mini when you have time. | 04:47.31 |
| Robin_Watts: are you around? I'd like to talk briefly about fz_begin_page() and fz_end_page() | 06:36.57 |
| Robin_Watts: I'm thinking that specialized implementations of fz_end_page() may throw and so they shouldn't be called in the clean up. in case e.g. fz_run_page() throws. | 06:39.06 |
| Robin_Watts: OTOH fz_begin_page() allocates a fz_device | 06:40.14 |
| Robin_Watts: I thinking that the device ought to be dropped in fz_drop_document_writer() unless an earlier call to fz_end_page() has dropped it. | 06:40.53 |
| Robin_Watts: to do so fz_drop_document_writer() would need to get the device and supply it to the specialized implementation. | 06:42.24 |
| I guess that makes sense. | 06:42.28 |
| https://android.googlesource.com/platform/bionic/+/master/android-changes-for-ndk-developers.md oh, interesting. Android recommends not linking to the platform libcrypto.so for crypto support as this will go away. | 07:25.26 |
| I guess this is another argument for doing signature validation in another way. | 07:31.45 |
sebras | foods. | 07:44.07 |
Cova | hi, I'm getting a n assembler error while compiling mupdf, like this: | 11:10.00 |
| {standard input}:458963: Error: can't resolve `.rodata' {.rodata section} - `.Ltext0' {.text section} | 11:10.12 |
| admittedly, I have non-standard setup (gentoo/lto/gcc 6.30), however, anyone has some hints on how to debug this? thanks | 11:11.26 |
| it is the lto-wrapper to fail, of course, but this seems to be linked to some assembler hiccups... | 11:12.31 |
| Forward 1 day (to 2017/06/12)>>> | |