View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0005012 | Taler | obsolete component | public | 2017-05-15 08:46 | 2018-04-15 18:34 |
| Reporter | Christian Grothoff | Assigned To | Marcello Stanisci | Resolution deadline | |
| Implementer | Reviewer | ||||
| Release signed by | Deployed by | ||||
| Verifier | Communicator | ||||
| Data protection authority notification required | No | ||||
| CSIRT notification required | No | ||||
| Estimated effort (hours) | Actual effort (hours) | 0 | |||
| Fast-track | No | ||||
| Work package | none | ||||
| Priority | normal | Severity | minor | Reproducibility | always |
| Status | closed | Resolution | fixed | ||
| Platform | i7 | OS | Debian GNU/Linux | OS Version | squeeze |
| Product Version | git (master) | ||||
| Target Version | 0.5 | Fixed in Version | 0.5 | ||
| Summary | 0005012: bank looks in wrong directory for default configuration | ||||
| Description | At least in the test_bank_api, it prints: 2017-05-15 08:47:43,430 talerconfig WARNING can't read config directory '/home/grothoff/lib/share/taler/config.d' which is wrong as "lib/" should not be there! Consequently, it then prints warnings like: 2017-05-15 08:47:43,432 talerconfig WARNING Configuration is missing option 'max_debt' in section 'bank', falling back to '50 KUDOS' 2017-05-15 08:47:43,432 talerconfig WARNING Configuration is missing option 'max_debt_bank' in section 'bank', falling back to '0 KUDOS' | ||||
| Steps To Reproduce | Note that I have TALER_PREFIX=/home/grothoff/lib (!) which _is_ allowed (as it might be lib64/) and the (GNUnet) spec for PREFIX says that the code must check for lib/lib32/lib64 and strip it of the prefix. | ||||
| Tags | No tags attached. | ||||
| Can't reproduce this.. | |
| Did you explicitly set your TALER_PREFIX environment variable to include /lib/? | |
| Now I did, and still can't reproduce. Here's the reason: in my prefix-end-with-lib/share/taler/config.d there actually is a bank.conf, so the config loader doesn't complain. In your case, you must have used some buggy 'make install' which didn't put any file in there. To answer your question: no, the config loader doesn't strip the final /lib part from the prefix. Anyways, I couldn't find those GNUnet specs you mentioned, so is the final '/lib' part to be removed only for the config's destination, or for everything (executables/libraries/etc.)? | |
| Well, for libraries obviously not, as they are in /lib/ (or /lib32/ or /lib64/, which is why we need the prefix to contain /lib/ in the first place). Executables that are found in /lib/libexec/ also need the /lib/. But for /share/ stuff, /lib/ needs to be stripped. | |
| Ok. As of libraries, I thought the right behaviour is to have --prefix=<x> and then 'make install' will install libraries in <x>/lib; you are never supposed to specify the final /lib in your prefix (unless you really want stuff to be in <x>/lib/lib), right? To recap: you ask that, whenever the user specifies a <x>/lib prefix, we strip the final /lib part from it with regard to /share/ stuff, and do NOT add a second /lib with regard to libraries and "libexec" executables. But what about executables (like taler-bank-manage)? I don't think we want them in <x>/lib/bin. Isn't it easier/more-obvious to just obey what the user gives as prefix and don't strip/add anything? | |
| Yes, --prefix=<X> will install to <X>/lib/. But if you set TALER_PREFIX (the environment variable), you _may_ set it to <X>/lib/ instead of just <X>. This is needed as on some platforms, the /lib/ ends up being /lib64/ or /lib32/ or /libGODKNOWSWHAT/, and so the TALER_PREFIX envrionment-variable override allows us to find out what GODKNOWSWHAT is. (Fortunately, we usually don't need TALER_PREFIX at all!) | |
| (Just to clarify: this is not about changing where the installation process installs files, it is about the way the Python code interprets the TALER_PREFIX environment variable!). | |
| Should be fixed by now (e22f7049d..). | |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2017-05-15 08:46 | Christian Grothoff | New Issue | |
| 2017-05-15 08:46 | Christian Grothoff | Status | new => assigned |
| 2017-05-15 08:46 | Christian Grothoff | Assigned To | => Marcello Stanisci |
| 2017-05-30 09:11 | Marcello Stanisci | Note Added: 0012192 | |
| 2017-05-30 09:25 | Christian Grothoff | Note Added: 0012193 | |
| 2017-10-23 08:55 | Christian Grothoff | Severity | major => minor |
| 2017-11-03 15:09 | Marcello Stanisci | Note Added: 0012540 | |
| 2017-11-03 19:46 | Christian Grothoff | Note Added: 0012545 | |
| 2017-11-04 10:24 | Marcello Stanisci | Note Added: 0012546 | |
| 2017-11-04 10:54 | Christian Grothoff | Note Added: 0012547 | |
| 2017-11-04 10:55 | Christian Grothoff | Note Added: 0012548 | |
| 2017-11-04 20:27 | Marcello Stanisci | Note Added: 0012564 | |
| 2017-11-04 20:27 | Marcello Stanisci | Status | assigned => informing |
| 2017-11-04 20:27 | Marcello Stanisci | Resolution | open => fixed |
| 2017-11-06 18:16 | Christian Grothoff | Fixed in Version | => 0.5 |
| 2017-11-06 18:16 | Christian Grothoff | Target Version | => 0.5 |
| 2018-04-15 18:34 | Christian Grothoff | Status | informing => closed |
| 2022-08-23 18:26 | Christian Grothoff | Category | bank (demonstrator) => py bank (demonstrator, obsolete) |
| 2023-12-03 00:23 | Christian Grothoff | Category | py bank (demonstrator, obsolete) => obsolete componet |
| 2023-12-11 19:08 | Florian Dold | Category | obsolete componet => obsolete component |