[sdsarchive] Search *.3 directories (mseed3)#181
[sdsarchive] Search *.3 directories (mseed3)#181andres-h wants to merge 1 commit intoSeisComP:mainfrom
Conversation
| // >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> | ||
| SDSArchive::SDSArchive() {} | ||
| SDSArchive::SDSArchive() { | ||
| _format = 0; |
There was a problem hiding this comment.
It is more readable to do an inline initialization at the member declaration:
int _format{0};
This also avoid initializing this member twice in different constructors.
| if ( name == "format" ) { | ||
| if ( !Core::fromString(_format, value) || | ||
| _format < 2 || _format > 3 ) | ||
| return false; |
There was a problem hiding this comment.
Logging an error would be nice, wouldn't it?
| bool SDSArchive::setSource(const string &src) { | ||
| if ( src.empty() ) { | ||
| string arcroots; | ||
| size_t pos = src.find('?'); |
There was a problem hiding this comment.
Here we could use Util::Url to parse the given source. This will give you access to URL parameters and do the proper decoding of all components.
| return resolveCha(pathStr, net, sta, loc, cha, requestStartTime, doy, year, first); | ||
| bool res = true; | ||
|
|
||
| if (_format == 0 || _format == 2) |
There was a problem hiding this comment.
Please add spaces within parentheses. Furthermore use curly brackets for conditional blocks even if they are just one-liners.
|
The code is in general looking good. Is this decision to use |
There is a pull request at EarthScope/slarchive#6 Anyway, it's just a proposal. I don't expect that slarchive4 fully replaces slarchive yet, but people who need mseed3 can use it together with seedlink4 and chain4_plugin from https://github.com/andres-h/seedlink4 I've heard that EarthScope is planning to start seedlink4 streaming in few months.
That would probably break too many tools. |
|
@andres-h, are you done with that PR? Do you consider it as final to be merged? |
|
I will make the requested changes. It is not ready yet. Furthermore, I think there may be problems with variable blocksize. I haven't tested it. |
|
I tested this with variable blocksize. It seems to work OK, albeit less efficient (linear search) and I get a warning like this: or more often this: Not sure if anything needs to be done. I created variable blocksize artificially and mseed3 with fixed blocksize (all blocks in file having the same size) works without any problems. |
|
@andres-h Can you quantify "less efficient"? |
3 participants
Not 100% sure if done correctly, but seems to work. Ideally mseed2 should take precedence if both formats are available.